Grid Dynamics Holdings Inc - Ordinary Shares - Class A Stock

Steinberg and CFO, Anil Doradla. Following the prepared remarks, we will open the call to your questions. Please note that today's conference call is being recorded.

Before we begin, I'd like to remind everyone that today's discussion will contain forward-looking statements. This includes our business and financial outlook and the answers to some of your questions. Such statements are subject to the risks and uncertainty as described in the company's earnings release and other filings with the SEC.

During this call, we will discuss certain non-GAAP measures of our performance. GAAP to non-GAAP financial reconciliations and supplemental financial information are provided in the earnings press release and the 8-K filed with the SEC. You can find all the information I just described in the Investor Relations section of our website.

I now turn the call over to Leonard, our CEO.

Thank you, Cary. Good afternoon, everyone, and thank you for joining us today. We delivered a solid second quarter, consolidated revenue of $108.2 million, above the high end of our guidance range and ahead of Wall Street expectations, with non-GAAP earnings of $14.7 million, which also is beating consensus.

As you may recall from my last quarter commentary, there were three areas I highlighted. First, improving revenue trends, especially with key accounts in the areas of technology and financial services; second, our AI adoption and growth; and third, improving profitability trends. I'm happy to report that on all three fronts, our execution is solid, and we're seeing the benefits, growing top account relationships, continued AI momentum with expanded capabilities in robotic and Physical AI and solid progress toward our 300 basis point margin expansion commitment.

For the second consecutive quarter, our top accounts are in technology and financial services. Technology and financial services now define our most strategic customer relationships and those precisely the sectors where AI adoption is moving fastest and where our capabilities are the most differentiated.

Our top accounts continue to drive our growth. Several delivered double-digit quarter-over-quarter growth with standout performances. There are no incremental gains, they reflect expanding programs, deeper program adoption and the compounding effect of our capabilities that keep finding the opportunity inside each client's organization.

Several of these clients are now embedding our GAIN platform as core infrastructure in their own operations, not just a project tool, but as a sustained capability. This is a fundamentally different and more durable commercial relationship than what we've had two years ago.

AI revenue reached 30.7% of the total company revenue in the second quarter, growing 54.6% year-over-year and crossing the 30% threshold for the first time. Two consecutive quarters of the year-over-year growth over 50% tells us something important. This is not a spike. It's a sustained shift. The trajectory is clear, and we intend to build on it.

Driving this strong performance is a combination of multiple factors. Our GAIN platforms are winning wider enterprise adoption. Our clients continue to transition enterprise AI workloads from pilots to production. Our engineers are more deeply embedded inside client organizations. Bottom line, we're winning entirely new programs that gives us confidence in growth ahead.

AI-first delivery is now the default, not the aspiration. Fixed price is a preferred approach on new RFP responses. The productivity and margin gains are real. We're executing well and delivery projects successfully. Our focus on executing larger AI platforms is aligned with significant progress we're making in upskilling our engineering talent. By the end of October, we plan to have 90% of our engineers trained on AI SDLC.

Our GAIN platforms have expanded LLM partnerships meaningfully this quarter. We're now working with several of the world's leading AI companies, including the top four frontier providers with whom we are under commercial agreements. This approach ensures our GAIN platforms stay aligned with the leading AI platforms with broader reach across our enterprise client base.

GAIN remains the backbone through which we bring AI capabilities to market. Its partner depth makes us stronger every quarter. Our client relationships are evolving too. Clients who came to us for platform deployments now ask us to stay. They want us to be involved in advisory execution ongoing operations. This meaningful shift is opening a growth vector that did not exist in our model two years ago.

On the partnership front, partner influence revenue reached 19.1% of the company total revenue in the second quarter. That was driven primarily by our three core hyperscaler relationships with Google Cloud, AWS and Microsoft Azure. A growing proportion of that revenue is coming from AI engagements. We are running Agentic AI workshops across our Google Vertex AI search customer base, converting search engagement into broader agentic commerce programs.

We extended our Google partnership in banking and financial services, closing our first joint win this quarter at a leading global bank. As we are deepening our AWS relationship around application modernization and Agentic AI in CPG manufacturing and financial services. Our NVIDIA partnership is gaining momentum across both Agentic AI and Physical AI. Our longer-term target remains 25% to 30% partner influence revenue, and we're confident of achieving this target.

Last quarter, I introduced our Physical AI capabilities and our first commercial engagements in the space. Physical AI requires a deep understanding of multiple disciplines that include modeling real-world robotics movements, digital twins, verification and simulators and integration with hardware systems. Our active programs span humanoid robotics for pharmaceutical intralogistics, autonomous driving stacks for construction equipment and policy control platforms for manufacturing clients.

We signed a strategic partnership with Doosan, a leading robotics manufacturer this quarter, elevating our NVIDIA relationship and opened an engineering office in Dresden, Germany to support our European manufacturing clients. Grid Dynamics enhanced robotics offering by welcoming Ekumen, a leading robotics engineering team that joined us in May. Their expertise resides in a Robot Operating System, a foundational open source standard that powers the vast majority of the world's industrial robots. Over the past decade, the company has built an invaluable list of some of the world's most respected robotics companies.

Grid Dynamics brings advanced AI modeling, policy control and enterprise scale delivery capability. Ekumen brings deep knowledge of the foundational software layer that robot manufacturers depend on. Together, the combination is formidable, spanning the full stack from the foundational software layer through simulation, hardware integration and enterprise scale deployment. We believe no other service company in the market today matches this combined footprint and technical depth.

Now let me pass on to Vasily Sizov, Chief Revenue Officer, who will expand on key business aspects of Grid Dynamics client engagements. Vasily?

Thank you, Leonard. Let me begin with three demand trends we observed during the quarter. First, clients are prioritizing AI investments that deliver clear, measurable business outcomes. Second, as clients move from isolated use cases to enterprise scale initiatives, they realize that the underlying technology layers must be modernized to support AI adoption. Third, clients increasingly recognize that successful AI transformation requires more than technology alone, driving interest in AI process consulting, performance benchmarking and change management. These trends align closely with our strategy and the capabilities we are building.

Let me discuss each of them in more detail. First, the demand environment remains constructive with clients directing AI investments toward practical application with tangible business impact. We are seeing particular interest in AI-enabled automation that improves operating efficiency, scalability and speed. Importantly, these investments are increasingly moving beyond experimentation with clients deploying AI capabilities into production to automate complex manual processes, improve customer service, reduce operating costs and create new sources of revenue.

Second, as clients move from isolated AI use cases toward enterprise scale transformation, they are finding that their data, application and core platforms must be modernized and made AI ready. As a result, AI adoption is creating broader demand across the underlying technology landscape. This trend aligns closely with our core expertise in data engineering, application modernization, cloud and platform engineering and reinforces the relevance of these capabilities in the era of AI.

Third, we are seeing growing demand for AI process consulting, performance benchmarking and change management as clients focus on converting AI investments into measurable business value. They need to identify the business processes where AI reengineering can create the greatest value, establish clear performance baselines, redesign those processes, build the technical enablers and drive enterprise-wide adoption.

We have been deliberately strengthening these capabilities to help clients realize measurable value from AI across the enterprise. These trends are reflected in our client work.

Let me highlight a few engagements from the quarter that demonstrate how these capabilities are being applied in practice. For a leading food service distribution company, we built and deployed an AI-powered product credit claims platform that automatically validates customer claims against photographic evidence. The platform cross-checks product, manufacturer label and shipping label images against the claims reason code in real time, replacing a fully manual salesperson-mediated review process.

In performance testing, the system processed approximately 400 claims supported by 1,000 images end-to-end in under 15 seconds per claim. The capability is now live in production, and the client has approved a long-term road map to further enhance the system and extend automated decision-making into more advanced credit adjudication scenarios.

For a leading home improvement retailer, Grid Dynamics enabled next-day delivery by designing and deploying a high load service that modernize the retailers' logistics operations. The solution includes an AI-powered routing capability that assigns fragile items to the appropriate vehicle types, eliminating hundreds of delivery errors each week. As a result, the solution cuts average delivery time by more than half from 3.5 days and is expected to support up to $0.5 billion in incremental annual revenue for the clients.

For a global technology company, we modernized large-scale data processing infrastructure, migrating more than 1,000 data pipelines to a serverless execution model. This reduced idle compute capacity, reduced infrastructure costs and improved scalability. Our proprietary AI-powered automation accelerated the migration and established a reusable delivery approach that is now being applied across broader initiatives at this client.

Now let me turn the call to Yury Gryzlov, our Chief Operating Officer.

Thank you, Vasily. Let me build on the Physical AI and robotics work Leonard introduced. Physical AI needs a full technology stack, and we operate across everything between the robot and the enterprise. The devices themselves come from our hardware partners.

At the foundation is the Robot Operating System, ROS and ROS 2, the open source layer the majority of the world's modern robots are built on, connecting the hardware to everything above it. Through acumen, we are not just users of it. We are among its maintainers and the founding member of the alliance that governs it. On the top of that sits the intelligence, the AI models that let a robot perceive its surroundings, generate its own motion and handle real-world variability. We design and validate that in simulation before it ever runs on a real robot.

And our own platform Incarnum, our GAIN platform for Physical AI is where enterprises bring it all together, building manipulation and inspection workflows, deploying those models and monitoring robotic lines with digital twins. What unifies it is our focus on the enterprise, expanding this capability to the companies that have robots deployed at scale. Here are a few examples that illustrate our work across the stack.

For a leading manufacturer of construction and mining equipment, we are building a next-generation stack for autonomous driving, loading and excavation. We are helping them design the platform, onboard the first use cases and add capabilities like policy-based control. What began as our first commercial Physical AI engagement is now a multiyear program across several regions.

With Ekumen, we've proven two-arm manipulation, grasping and assembly trained entirely in simulation and then run reliably on a real robot. Bridging that gap from simulation to the physical robot is one of the hardest problems in the field. Humanoids are the next step. A leading life sciences company is piloting humanoid robots for intralogistics, moving and repacking containers of chemicals, work that was out of reach only a couple of years ago and is now possible, thanks to new AI models that generate motion. We provide the platform those robots run on working with Wandelbots and on NVIDIA stack. The customer calls it a lighthouse project for their industry, and it's the opening step in a much wider program.

We are also building the channels to scale. This quarter, we announced a strategic partnership with Doosan Robotics, a global leader in collaborative robots deployed across 45 countries. It's a full stack collaboration. Our platform plus the foundational AI components, integration services and engineering around it paired with Doosan's cobots and our combined global reach.

Together, we can provide what traditional robotic software can't, dual alarm assembly, inspection of complex geometry parts and packing of deformable items. It sits alongside our elevated NVIDIA partnership, and we are in active talks with several more hardware and software vendors. Considering the economics of software services in this space and our positioning, we are confident that we have a material market advantage.

Reliable performance in the physical world takes engineers who understand simulation, control and hardware variability, working through problems that have no templated solution and so can't be easily automated. This combination is hard to assemble. Ekumen's decade of foundational robotics depth together with our strength in AI modeling, simulation and enterprise delivery.

We don't believe another services company matches it today. Closing that gap isn't a matter of hiring a team. It's years of hard-won experience, which we are now putting to work for our customers.

In summary, robotics and Physical AI is a growing market measured in the trillions over the coming decade. Our expanded capability is helping us capitalize on the early traction we saw last year reflected in a rapidly growing pipeline from both existing customers and new logos.

Another important part of my update is tied to our capital markets focus, where a similar pattern is playing out in software rather than robots. As our banking clients push Agentic AI deep into their engineering, the hard part is no longer producing code, it's doing it safely with quality, security and control they can provide to a regulator. This quarter, that showed up most sharply around security. Banks want the speed of frontier models and AI-generated code without introducing new vulnerabilities. Our answer is spec-driven Agentic engineering led by Allium, part of our GAIN platform for AI SDLC and is exhibiting real traction across our banking clients.

The clearest example is at one of the world's largest banks where Allium is being used to build new tools as part of a bank-wide initiative to modernize business operations. Working across London, New York and India, we are bringing specification-driven development to both new and existing systems, starting with tools for AI-assisted productivity and extending to agents that automate operational work. Taken together, Physical AI reaching the enterprise and AI native engineering scaling inside the world's largest banks, this is the frontier work that keeps Grid Dynamics differentiated. Over to you, Eugene.

Thank you, Yury. Good afternoon. Last year, I described our AI strategy through three horizons. This quarter, I'll describe them by maturity. What has reached scale and what is beginning to scale. Horizon One, scaled, AI first modernization and the agentic platform. Modernization remains the foundation of our business, but AI is changing how the work gets done. Agents can now accelerate work across most of the modernization life cycle, particularly code generation and testing. The remaining work, business acceptance, production scaling and complex consummation still depends on human judgment and accountability.

An agent can write a code, a person still makes a call and stands behind it. We have invested in a set of GAIN tools that support this life cycle. Rosetta governs how agents operate. Allium analyzes legacy systems to create reliable specifications for their replacements. And SpecFlow, our latest open source contribution uses those specifications to support autonomous feature implementation.

Rosetta has progressed from its first lighthouse clients to larger engagements across retail, financial services and manufacturing. At a Fortune 30 U.S. home improvement retailer, approximately 550 of our clients' engineers are working with the platform. In one program, seven COBOL services were moved to a modern technology stack with approximately 90% of the code generated by agents. All seven services entered production this quarter.

The client already had capable engineers and access to many of the same AI tools we use. What it needed from us was domain knowledge, governance and control, the capabilities between powerful agents in the dependable enterprise systems. This productivity is helping us expand client relationships. It is also creating opportunities to use more fixed price and outcome-based commercial models when the scope and accountability are clearly defined.

Allium also reached an important milestone this quarter. It has been piloted across five major banks and has begun moving into its first commercial banking engagements. Allium analyzes legacy code to help establish reliable functional specifications for replacement systems. It also supports controlled migration and rollback, reducing the operational risk of moving critical applications on to modern platforms.

At one major North American bank and one-hour GAIN demonstration in February led to a signed contract in April. The bank was managing 150 applications with limited test coverage and a ground security backlog. We translated identified issues into failing tests inside the bank's own touring, allowing its engineers to independently reproduce and assess each finding. At another Tier 1 bank, the same approach is supporting a security modernization program spanning more than 100,000 systems. This part of the modernization work co-founded by the client's cloud provider.

Our differentiation is not limited to code generation. Our agents can also incorporate context such as security advisories, dependencies and upstream changes. That broader context helps identify problems that code-only tools can miss and provides the traceability and evidence regulated enterprises expect.

We deliberately make selected GAIN platforms open source. The immediate objective is adoption and technical credibility, not software license revenue. Open code allows engineering leaders to evaluate our capabilities directly and strengthen our position when client needs help deploying those capabilities and enterprise scale. The same pattern applies to data. Enterprise AI cannot deliver reliable results without accessible well-governed data. That is increasing demand for data platform modernization. Our new AI data migration accelerator released this quarter is already being deployed in data lake modernization program for global consumer products manufacturing.

The second scale component of Horizon One is GAIN Agentic Runtime. Enterprise agents need access to trusted data, evidence that their behavior is controlled and governance over operating costs. For a global payment client, we brought these capabilities together as shared services with retrieval layer now supporting 25 enterprise consumers. We also converted the client's dispute architecture, including fraud, chargebacks and KYC to configuration-driven workloads.

A common foundation now supports four use cases. By automating much of this configuration, the program rebuilt a decade of business logic in just six months and reduced integration and release cycle times by 96%. At our largest banking client, an internal platform built with our support now centralizes the registration, governance and operation of AI agents across the organization. The client reports regular adoption by more than 80% of its employees across more than 80 markets. As adoption grows, we are also developing the operational tooling needed to govern and support the platform at that scale. Across these engagements, the pattern is consistent. AI accelerate production, but enterprise value comes from the domain knowledge, governance and accountability required to put it all into production responsibly.

Horizon Two, scaling. Harness engineering and Physical AI. Horizon Two covers capabilities that are moving from research and internal validation towards repeatable client deployment. The first is agentic harness engineering. Traditional agentic workflows are most effective when the task and sequence of step are already known. Harnesses are designed for more dynamic work, situations in which an agent must select tools, adjust its approach and respond to new information while remaining with defined controls.

The harness provides those controls. It records what the agent need, test its output, manages exceptions and introduces human review where accountability requires it. This allows enterprises to apply agents to more complex work without giving up oversight. During the second quarter, our engineering center developed nine harness-based solutions. Following our client zero approach, we are testing them first with our own operations. The objective is to establish evidence of reliability, define the necessary controls and improve the solutions before introducing them into client environments.

The second area is Physical AI and robotics. We are investing here because the engineering challenge is fundamentally different from conventional software development. A coding agent can generate software and test it in digital environment. A physical system must also operate safely and reliably in the real world. It must account for geometry, motion, changing condition and the behavior of physical environment.

Validation, therefore, has to take place both in simulation and in hardware. A language model alone cannot close this loop. Our research is focused on bringing physics, geometry, simulation and continuous validation into the agent's operating environment. That is also the strategic rationale for the robotics engineering team we acquired in May. Members of this team have long contributed to core infrastructure in the Robotics Operating System ecosystem with particular expertise in simulation and validation. Their capabilities are now contributing to gain for Physical AI, our platform built on Incarnum.

During the quarter, we released three new components, tools for composing robotic policies, a continuous improvement loop and Sandbox environment for control testing. We are beginning to validate the platform through early client and partner deployments. A leading life cycle company is piloting humanoid robots in its warehouse operations using our platform. Separately, a robotics partner has incorporated the platform into its own offering, creating a distribution channel for our Physical AI technology.

Horizon Two is not yet the same maturity as our modernization and Agentic platform business. Our focus now is to demonstrate repeatability, convert technical validation into production deployments and establish scalable commercial models. The opportunity is to build a differentiated intellectual property in areas where success requires not only generating software, but providing how the software behaves in the physical world.

Across both horizons, the pattern is clear. The cost of producing software is falling, while the value of governing it, validating it and taking responsibility for it in production is increasing. This quarter, more components of gain moved from tools and pilots into broader enterprise adoption. At the same time, our investments in Agentic harnesses and Physical AI progressed from research towards controlled client deployments. As these capabilities mature, they allow us to reuse more of our engineering, deploy solutions faster and take greater responsibility for measurable outcomes.

Our advantage is not simply that our agents can generate code. It is that they combine those agents with domain knowledge, operational controls and the engineering discipline required to make them dependable at enterprise scale. That is where we believe durable value will be created the agentic era and where Grid Dynamics is positioned to lead. Anil, over to you.

Thanks, Eugene. Good afternoon, everyone. Second quarter came in at $108.2 million, slightly above the higher end of our guidance range of $106 million to $108 million. That represents 7% year-over-year growth, including de minimis contributions from Ekumen. Non-GAAP EBITDA was $14.7 million or 13.6% of revenues and was closer to the high end of our $14 million to $15 million guidance range.

Looking at the performance of our verticals, TMT remained our largest vertical and accounted for 31.8% of total revenues for the quarter with a growth of 11.7% sequentially and 36.4% on a year-over-year basis. The growth was primarily driven by our largest technology customers. We continue to benefit from vendor consolidation at these customers, which has driven increased wallet share across new and existing programs.

Retail contributed 26.5% of total revenues in the second quarter of 2026. The vertical was flat in absolute dollars on a year-over-year basis and grew 3.1% sequentially. The sequential growth was supported by demand from key accounts, including a major specialty retailer.

Our finance vertical accounted for 22.9% of total revenues in the quarter and grew 1.2% on a sequential basis. Within this vertical, we witnessed solid demand from our fintech service engagements, including increased contributions from a major payments network, which helped offset the successful completion of engagements with insurance and data analytics and consumer credit reporting clients in North America. Looking ahead to the remainder of 2026, we remain bullish on our growth outlook within this vertical.

CPG and manufacturing represented 10.9% of quarterly revenues and grew 2.1% on a sequential basis and 4.2% on a year-over-year basis. Within this vertical, we are witnessing robust demand from a leading wholesale food distributor, along with growth from some of our manufacturing customers.

Turning to our remaining verticals. Our other vertical contributed 6% of our second quarter revenues, while health care and pharma contributed for 1.9% of our revenues for the quarter. We ended the second quarter with a total headcount of 4,838, down from 4,964 employees in the first quarter of 2026 and from 5,013 in the second quarter of 2025. We continue to rationalize our overall headcount as well as align our skill sets and geographic mix.

At the end of the second quarter of 2026, our total U.S. headcount was 379 or 7.8% of our company's total headcount versus 7.2% in the year ago quarter. Our non-U.S. headcount located in Europe, Americas and India was 4,459 or 92.2%. In the second quarter, revenues from our top five and top 10 customers were 43.5% and 61.5%, respectively, versus 37.5% and 57.3% in the same period a year ago, respectively.

Moving to the income statement. Our GAAP gross profit during the quarter was $39.6 million or 36.6% compared to $36.2 million or 34.8% in the first quarter of 2026 and $34.5 million or 34.1% in the year ago quarter. On a non-GAAP basis, our gross profit was $40 million or 36.9% compared to $36.7 million or 35.3% in the first quarter of 2026 and $35.1 million or 34.7% in the year ago quarter.

On a year-over-year basis, the increase in the gross margin percentage was primarily driven by revenue growth outpacing delivery cost. On a sequential basis, the increase in gross margin percentage was due to a combination of working time and improved resource utilization. Non-GAAP EBITDA during the second quarter that excluded interest income, expenses, provisions for income taxes, depreciation and amortization, stock-based compensation, restructuring, expenses related to geographic reorganization and transaction and other related costs was $14.7 million or 13.6% of revenues versus $12.5 million or 12% of revenues in the first quarter of 2026 and was up from $12.7 million or 12.6% in the year ago quarter. The sequential and year-over-year growth in EBITDA was largely due to a combination of higher revenues and strong operating leverage across our non-engineering overhead.

Our GAAP net income in the second quarter was $2.9 million or $0.03 per share based on a diluted share count of 83 million shares compared to the first quarter net loss of $1.5 million or a loss of $0.02 per share based on a diluted share count of 84.7 million and net income of $5.3 million or $0.06 per share based on 86.4 million diluted shares in the year ago quarter.

On a non-GAAP basis, in the second quarter, our non-GAAP net income was $9 million or $0.11 per share based on 83 million diluted shares compared to the first quarter non-GAAP net income of $7.5 million or $0.09 per share based on 85.9 million diluted shares and $8.3 million or $0.10 per share based on 86.4 million diluted shares in the year ago quarter.

On June 30, 2026, our cash and cash equivalents totaled $298.4 million, down from $327.5 million on March 31, 2026. Since our first quarter earnings call, we've repurchased approximately 2.6 million shares for a total consideration of $17.3 million. Cumulatively, since our Board authorized a $50 million share repurchase program, we have repurchased approximately 4.4 million shares for a total of $30.8 million, reflecting our continued confidence in the long-term value of the business.

Coming to the third quarter guidance, we expect revenues to be in the range of $112 million to $114 million. We expect our third quarter non-GAAP EBITDA to be in the range of $16.5 million to $17.5 million. For the third quarter, we expect our basic share count to be in the range of 81 million to 82 million shares and our diluted share count to be in the range of 83 million to 84 million shares.

For 2026, we're maintaining our full year revenue outlook of $435 million to $465 million. That concludes my prepared remarks. We are now ready to take questions. Cary?

[Audio Gap]

[Operator Instructions] Thank you, Anil. [Operator Instructions] The first question today comes from Mayank Tandon of Needham.

Congrats on the quarter, Leonard and Anil. So there was a lot of detail around AI. So just to step back, Leonard, could you maybe talk about the AI efforts and the implications for both growth and profitability over the next, say, 12, 24 months? Maybe you can help reassure investors that AI will actually be a net positive for you because there's still a lot of skeptics out there that I think is going to be a net negative over time.

Right. Thank you, Mayank. Well, it's a pretty comprehensive question. And if I answer all of the parts, there will be probably nothing left for the other. So I'll try to be concise in terms of the key elements, and then we can talk a little bit more in detail.

So first of all, we are reaching many aspects of AI implementations. We talked about it in the past, we're adding those features now. We're talking about directly or indirectly about forward deployed engineers. We make announcements. We train a substantial number of the people in the workforce. And these people are basically driving a new way of implementing our solutions because as we tend to get more focused on fixed bid and fixed budget projects, it helps us to identify not only the execution of the various modernization projects, but also create a technology consulting. So that's with respect of the people and why it's accretive to us.

When it comes to Agentic AI and a part of the, again, implementation of the suite of our solutions, we are driving our customers to adopt our GAIN platform model. All the elements of the model are driven by internal tested and development, but also tailored to our customer needs. So they will need to adapt the solution where the -- they see the most fee for themselves, but also we guide them through the process to create the best ROI for that.

So that's the second part. And before I talk about the physical, I want to bring -- to address your point in terms of net positive versus net negative. If you look at the increased growth in just these two areas, that substantially exceeds the -- some of the aged businesses, which would eventually drop out because the gloom and doom from many facets were about that engineering and consultancy is less relevant.

Moreover, people would say it's easier to train FTEs. We embrace FTEs. We embrace our clients. At the same time, as many of the leaders in interest is saying, we can do more work. We can do more engagements, which we prove with all the listed examples. And I'm not going to go through all of that because we have a lot of people who can give you more details on that.

So as a consolidated effort as we go today through further discussions, we will demonstrate on specific examples where this accretiveness works. But I want to emphasize forward deploy engineering and Agentic AI.

And the third part, which is also super critical for us, which actually drives the adoption and partnership enhancement of our relationship to the next level is actually our preparation for Physical AI work. We not just made a small acquisition. We not just made announcement about opening additional robotics labs. We're working -- we've been working with our clients for a long enough time to understand what it means for their own platform, what it means for their application and solutions from various world from industrial, from modern machineries to logistics companies to even work in industrialization of various new solutions.

So the material side, the remuneration for the Physical AI is still to come. But now we have an evidence of substantial more players looking at Grid Dynamics again in a leadership role by expanding our capabilities to the practical world of their usage. So this is pretty much a summary and of course, we'll go in more detail.

That's very helpful. And sorry, if you can't see me, I'm having an issue with my video, maybe I can help with that. So I'll try to get that fixed eventually. And then just a very quick follow-up, Anil, for you. In terms of the guide, I just want to get a sense of the visibility that you have today versus last quarter. And what I mean by that is, is the pipeline now converting faster? Have you seen evidence of that? Does that maybe give you more confidence in the sustainability of growth acceleration once we get beyond fiscal '26 into fiscal '27?

You're talking about next year. Well, let's talk about this year and then we'll get to next year. When you go into the second half, Mayank, you see -- if you look at our visibility and our second half, there are a couple of factors. Number one, remember, the 85/10/5, most of our revenue comes from customers who have been with us for two years and beyond. That's formula more or less stays well intact. And that you're seeing in the top five, top 10 customers, right, because most of the absolute dollar and year-over-year growth is coming there. So that stays intact.

As you go into the second half, there are three layers as you go. First is the working time. Second half is higher than the first half. Second thing is that the billable headcount. So we're seeing new programs kicking in. So maybe without addressing your pipeline question directly, indirectly is that, yes, we're seeing an increased billable headcount as we go into the second half. And the third thing is that we are planning some acquisitions. So all these three add up to layers.

Now when you look into 2027, I think I'll let the business guys chime in here. But from my point of view, I see two things that are very interesting. Number one, the relationships that we're having with our technology customers, our financial customers, our top 10 and 20 customers is going deeper and deeper. Things that we've not done, we're doing. Application modernization programs, which we've not done, we're addressing. The addressable market that we're going after is larger. And I overhear these conversations week after week, which leads me to believe as you go into 2027, if we continue winning at the rate that we're winning, it should play out incrementally positive. But I don't know, Vasily or Yury, whether you want to add anything to that?

Yes. Let me chime in. So yes, I would say that our position with most of our biggest clients is -- has been strengthening over the last few years by -- through vendor consolidation. And what we see is that we should benefit in the coming years from this consolidation, which means bigger programs would come our way by customers cutting lose the long tail of vendors, which are no longer relevant. And given our strong technology positioning in Agentic AI, which is a very, very hot topic for most of our customers, we are really well positioned to benefit from that.

The next question come from Bryan Bergin, TD Securities.

So maybe just to start, a follow-up on that last question as it relates to kind of that second half, more of a near-term question. Just as it relates to you gave us 3Q guide, implied 4Q is still a decent ramp. Are you seeing a broadening of momentum in other sectors? Obviously, doing quite well in technology. Are you seeing a broadening of momentum elsewhere that gives you that confidence?

And then as it relates to kind of potentially some M&A requirements, any way you can share with us how you're thinking about maybe the organic contribution remaining versus any needed M&A that you have to go get?

Right. So Bryan, let me point out that, as you know, there's a certain seasonality in our business, right, as we go into Q3, Q4, that's established. As I said, there are three levels at which we're operating. Number one is just the working times of the second half of the year, and you guys know it's better. Second thing is that the billable headcount and the trends are positive and all our prepared commentary should lead you to conclude that.

And the third thing is that there is a certain amount of acquisition. And we do have a pipeline. It varies -- I always joke, right? An acquisition is not done until the money is transferred to their bank, right? We've seen acquisitions that we thought are not going to happen -- they happen. We've seen acquisitions that were locked and loaded, and we just are not able to.

So -- if you look at that second half, I don't want to comment too much on Q4 other than saying that, look, we have a seasonal pattern for the year. But as we go from the low end of our full year guide to the high end of the guide, the first component of working time stays intact. The second component of billable headcount, we have variable calculations. And the third component perhaps picks up a little bit more is the acquisitions.

Okay. Understood. My follow-up is kind of a margin and a tie-in with the delivery model question. So it's -- you reiterated the confidence in the 300 basis point expansion, so that's good to hear. I'm just curious how much of this margin improvement is coming from structural changes, automation and efficiencies in the delivery versus traditional kind of cost control cutting measures. And I think it's notable you had like 7% revenue growth, while headcount was down 3%. Is there a lasting -- I know you're saying you're going to add double headcount, but is there a lasting change in this delivery model? So just maybe talk about that AI-driven efficiency and delivery that you're seeing.

So there are three, four parts of this question. Let me take the first part and then when it comes to some of the AI trends, I'll pass it on. But when you look at what we set out to do, we said that on a year-over-year, we're going to deliver 300 bps margins on a Q4 by Q4 on a year-over-year basis.

Part of that effort is efficiency is just the way we're organized. As you know, we've ramped from a handful of countries to 19 countries. We've got many incorporated entities. So there's a little bit of an efficiency that we brought in, and some of those are onetime, but we operate at a certain level, right?

The second part that we are seeing here is we're embracing a little bit more change in the way we're doing business, whether it's AI, whether it's fixed price, whether it's a greater -- embracing more tools. And that is creating a certain level of -- I would say it's not so visible now, but over time, you'll see a non-linearity perhaps that is in.

The movement that you've seen on the headcount right now was largely driven by efficiency improvements on non-engineering headcount. So people should not worry, it's not that we let go some billable headcount. No, it's just non-engineering, non-billable headcount, we cleaned it up. But from this point onwards, beyond the 300 basis points that you'll have from Q4 to Q4, as you go into 2027, there is a plan for us to leverage more of these tools. There is a plan of bringing a certain level of non-linearity. We have the plans. The clients have to accept it, and we have to proceed with that. Go ahead, Leonard.

Yes. Let me add a couple of things. First of all, just to complete answer on the very first question about diversification of the platforms. I think it's very critical to understand that this is not overnight we selling diversified verticals. First and foremost, we've been in the payment system, we've been in financial services, we've been industrial and modernization.

The second of all is what you can actually see from the previous comments about us what Vasily said we're replacing some incumbent vendors is because we're playing in a big boys league, in a higher level. I mean, in the past, look, there were always couple top guys and there were a couple mid-level vendors. Now we only compete with the top guys.

And the reason being is I think AI adoption and technology implementation equalize the field a bit. So we've always been prepared for the big tasks and a big program, big transformational solutions, but we also gained the reputation of this consultancy part. So as we get more admittance to the bigger projects, inevitably, what happened with that, it's a better visibility, better projection, better position.

So we're saving with the tools. We're adding more capabilities. And we're looking back and we say, what of these internal systems, which we had for a long time are less efficient. We accepted to live on a world of uncertainty. That's very important. We don't see the world changing so dramatically. We'll go back immediately to the luxury of being very consolidated in very few locations. We're adding not only India, and we're adding investment into India and the technology capability, but also LatAm. And we -- as we do more, we create a global platform internally to optimize this efficiency. So it's a cost structure, it's performance-based, it's tooling, it's removing redundancies from the past. I hope, Bryan, I covered a lot.

The next question comes from Puneet Jain of JPMorgan.

So how are your AI and robotics partnership different from your traditional hyperscaler relationships like with Google, AWS, Microsoft Azure that generate much of your 19% of partnership revenue? So the partnerships you got with NVIDIA model companies, do they differ -- or do they offer like a different revenue trajectory potential or client ownership structure than your other partnerships?

All right. Thank you so much for the question, Puneet. Let me address this question. So we definitely value our relationships with NVIDIA and believe that's a great partnership to build a pipeline of future opportunities on. As you understand, right now, the industry, the manufacturing is going through a massive transformation and new tools like Agentic AI or Physical AI definitely brings new technology to more like a traditional manufacturing. And we see this as a great opportunity to build a new pipeline of opportunities and new type of engagements, which would help us to transform those manufacturers on a bigger scale.

So just an example. So for example, right now, we have an active engagement with one of the world's largest industrial equipment manufacturer on building a Agentic AI platform, which allows to manage the fleet of autonomous vehicles and deploy Physical AI capabilities on the edge devices. And we see more and more interest to such opportunities. So it's definitely one of the top priorities for us to grow.

Got it. And I'd like to follow up on the prior question, specifically around headcount. I noticed like your non-U.S. headcount was down despite like the Ekumen, which probably contributed employees in Argentina. And the U.S. headcount by comparison was up on a sequential basis. So should we expect this remix to continue like as you do more AI-based services, will that require more on-site headcount or U.S. headcount compared to in the past? And if that's true, what does that mean for margin and change management within your employee base?

Very good. Well, Puneet, what you said, it's music to Eugene's ear because he's been the one who is architecting the acceleration of some of the U.S.-based presence, both from the technology office perspective, but also from the consultancy -- technology consultancy with the clients.

I'm not saying there's more shift towards onshoring as a trend. I think if you look back pre-COVID days, our onshore presence between onshore technology people as well as some of the offshoring engineers who would come on the long-term projects reached almost close to 20%. It's never been so low. And then when the onshoring presence pulled back due to the -- an ability to work directly with the clients, a lot of work has been going on offshoring.

Now we're not saying that work is no longer relevant. But there are more and more demand presence on-premise with the clients to work together on these complex cases because the rapid change of transformation sometimes catches the clients a little bit through uncertainty, right?

We talk about two basic approaches to their mental and budgetary resolution of the projects. One of them is more like a status quo. Let's see and tell what's going to happen. Obviously, they don't need as much of onshoring presence. And some of them demand very rapid acceleration, but they are concerned of some of the spendings, as you know, around tokens and other things, which definitely create the pressure. So that's where our headcount onshoring technology-wise is coming.

As I mentioned to Bryan, some of the reduction of offshoring headcount comes from non-engineering and non, I would say, forward-looking specialties. So there is a difference between the headcount and contribution of this headcount. So from the budget perspective, it's a little bit less clear that these people were extremely expensive, but just the infrastructure on these people would no longer be needed for us to serve the markets better.

So to answer your question, we do see some additional growth of onshoring. The ability of us to prove that our margin expansion will continue to grow is vastly driven how much of the fixed bid, fixed budget projects we can adopt, how much of our internal developed tools are accepted by the clients, how much of the non-linear value we're bringing to the party. And I think we're quite growing with those elements.

So just to conclude on that from my side and if people want to add, I think you picked the right trend. I don't think the legacy, some of the people are a sign for concern because majority of them come from the Central Eastern Europe. And I think this is all by the book. So we are really moving forward with a clear plan on continue to have margin improvement.

The next question comes from Matt Dezort of William Blair.

Congrats on the results. I wanted to see if you could double-click on this new consultancy practice that you're talking about. Can you discuss more of how you see this business developing? I know you talked about activities like change management, but what sort of opportunities are you seeing in the pipeline build there? Who are you going up against in these bake-offs? And how is the competitive environment different from your traditional work maybe?

Yes. So I will start very briefly and then Vasily will actually expand on it. So there are two parts of it. The first part is there is no change of our purpose. Consultancy has always been a part of our DNA. Nothing is like earth-shattering because our clients consider us to be a technology consultants. And that's why we're able to compete against the big firms.

What has changed is the distribution of that kind of offering. And just the previous question with Puneet was about onshoring presence, right? And people who we hire, they're extremely technical, but they're also customer-oriented. And that kind of work very important because we are expanding the purpose of consultancy from pure technology consultancies and now adding AI infrastructure consultancy, harness selection consultancy, tool selection consultancy and to some extent, getting more into the sacred world of business consultancy. So Vasily.

Yes. So think about business consultancy as a natural extension of our technology enabler build-out capabilities. Essentially, the focus of the customers is shifting from just creation of the system, but creation of the systems to change business processes they have. Therefore, they would like to analyze first which business processes are the best candidates to improve, which value is hidden there, then to build a technical enabler to reveal this value and then adopt that technical enabler on the enterprise wide scale. And that's exactly where the focus of our consultancy is not only to create the technical enabler, but also to help get all the value on the enterprise scale from this change. So that's the essence.

And right now, we have several active engagements on that specifically on the front of consulting change management, which goes along with technical enablers. And we see this opportunity for a great growth in the future.

I can add to that, that many of our customers observe the performance and productivity of our delivery teams using our GAIN platforms. They become interested, and they want those platforms and those methodologies inside their own software factory. And we are helping them to establish the tools, methodology and change management, which is required to gain similar productivities in the broader organization.

That's a good segue, Eugene, for my follow-up on GAIN adoption and just the S-curve that implies. As you accelerate GAIN rollout, how should we think about that adoption curve in AI -- pure AI revenue? Is it likely to scale linearly? Or you're talking about wallet share gains from AI. Is there a way we could see some exponential growth? And how could you drive more sharper inflection in that AI penetration with GAIN?

So what is interesting about our GAIN strategy is that we are consolidating all our IP from multiple accounts, from multiple practices under the same umbrella. And AI helps us to do that very, very rapidly and quickly. And our embedded engineers, forward deployed engineers are all tasked to bring back the learnings, the ideas, the -- what works and what not works back to the GAIN platform. And part of the GAIN platform is also open source that helps to drive the insights from the broader community and put the GAIN platforms in front of many leaders.

At this point in time, we observe a growth of having direct revenue from GAIN platform, but much more importantly, observe the growth of the overall connection and expansion of our relationships inside our accounts and the new accounts, which are driven by those platforms. So we see many of the inbound interest and conversations, which result in the new leads, new opportunities and new converted business from GAIN platform. This is what is happening right now.

Our next question comes from Surinder Thind of Jefferies.

I'd like to start with a question just around this idea of there's a bit more excitement around moving from proof of concept to maybe the actual implementation projects. And that commentary seems to be a bit more universal. From your perspective, can you maybe talk about what the revenue journey for that looks like, meaning how big a proof-of-concept project would be if it's a few hundred thousand dollars? Does that become a $2 million project? Or what's kind of the range of outcomes that we can expect here as we think about more of those proof of concepts coming and how that would impact the growth rate?

Surinder -- so again, I will give you a little bit of a high level, and I think because it's all revenue touched, Vasily will give you a little bit more color. So there are different proof of concepts. The definition of proof of concept could be quite stretched, both from intent and dollars associated with that and follow-ups.

When we looked at proof of concepts as a result of our partnerships, for example, that resulted in some of the very meaningful programs where the customers embrace not only our partner solution, but our offering, which was in conjunction with these partnerships. When we look today and specifically at the suite of GAIN productivity, it's actually very interesting.

Eugene mentioned about the inbound interest. As you know, for us, for Grid Dynamics and our size and capabilities, visibility is very critical. So the customers would reach to us with something we still call proof of concept, but it was a substantial projects because the measurement of proof of concept is sometimes driven today not by the amount of dollars could be quite more substantial in many cases, but the time to implement.

The whole short-term engagement definition, which used to be followed or preceded by the proof of concept, become the proof of concept itself, and then it's a major rollout. So this has conceptually changed the definition of proof of concept and revenue associated with it. But I'm sure that Vasily will give some more details.

Yes. So many customers start definitely with implementation of some smaller pieces of business cases, which have tangible business results in order to demonstrate it for their Board, for their management. And then using that as an example, essentially request more investment into that, which eventually gets converted into platform build-out -- to build AI harness and et cetera. And this trend definitely persists. That's what we see with our customers.

And I can say that for platforms work, this work essentially is much more sticky and longer term in nature than the POCs. And having built the platform, of course, there is a growing appetite to build more and more business cases on top of this platform. So it grows like a snowball, and that's actually is what's reflected in our pipeline.

And I think I just wanted to add that it also depends on the industry, right? We mentioned today about Physical AI and robotics. Definitely, there's a lot of proof of concept in those areas. But at the same time, it also depends on the -- how deep you are in your relationship with the customer and that other programs around outside even of those areas, right? And that's where those proof of concepts could be actually quite significant. Sometimes it could be just maybe a few weeks of small engagement. Sometimes it could be six months plus. And going back to the GAIN model and our platforms in the GAIN, I think this is where also we try to condense this knowledge, right, in a way to speed up this implementation as much as possible for our customer. And that also contributes to the ratio of the proof-of-concept revenue versus the longer engagement implementation revenue.

So just to summarize for Surinder, the POCs associated with FDE consultancy, GAIN model modernization subjects and Agentic AI as overall are very substantial from get-go. The Physical AI part is what traditionally we would call proof of concept because it's such an innovative way to modernize modern productivity and interface between human robotics. So these type of POCs are more traditional way and their revenue will follow with the scale which you typically expect from POCs.

Helpful. And then maybe thinking about the data and AI practice and the really high growth rate that we're seeing there, the 30% of revenues. Can you help me understand what's going on in the upper 70%? When I do the math, I get to roughly about a 10% decline in that other 70% of revenues. Is that -- how much of that is just cannibalization by the data and AI component? Because I assume every new piece of work probably falls into that bucket.

And then is there components that are maybe in that legacy, I'll call it, legacy bucket for lack of a better word, other elements to that, such as pricing compression or just other factors to think about?

Very good question, Surinder. I'm actually going to make it much simpler. It's not even that complex. In our case, when you look at our business trends, from time to time, we might see a significant customer. I mean, I'm loosely using that word, maybe a top 30 customer, top 40 customer, right -- have some changes. Maybe there's a change in strategy or a big project is completed or something changes, and we can have some volatility there. If you go back over the past couple of quarters and look back at Leonard's commentary, he talked about sensitivity of brick-and-mortar retail, for example. And you -- create some volatilities there. So very simply put, if I were to extract some of these volatilities there, we would see a better growth pattern.

Second point is your question is whether there is cannibalization. What we see is -- and Eugene and Vasily can back me up on this --when we get into our clients, especially in our top 20 clients, we're going deeper and deeper. So all the work is incremental AI work that we typically see.

And the third thing that we see here is on the pricing, we are not seeing pricing pressures. As a matter of fact, when you go into the AI world, obviously, there's a premium. But when you look at our -- what we are doing over the past couple of years and I look at a certain grade in a certain country and whether there's pricing pressures, the answer is absolutely no. We're not seeing that. Now we can argue whether there's a pricing increase, that's a different story, but there's no pricing declines.

Vasily, I don't know whether you want to add.

Yes. So I think it's a question of semantics, right? The cost per project or cost per functionality or piece of scope is definitely getting reduced because of the higher productivity and shortening the timelines for the delivery. That's kind of what's happening. But that leads to more work and more projects rather than the reduction. So I think that's a very important color.

And finally, you can look at the revenue per person. Again, we need to have some history to prove that that trend will continue to grow. So we're not trying to defend legacy. I think Anil was quite clear that some business just fall off, right? And you're absolutely right. The AI content brings more new business. But there is one element which is not there is us trying to retain some legacy business and compressing our margin. That's just not part of it.

I mean it does sound like there's a definitional component here, right, of how you -- to your earlier point of how you divide up the buckets and trying to look at it collectively plus all of the kind of the noise of project starts and stops and things like that. So I appreciate that.

Ladies and gentlemen, this concludes our Q&A session for today. I will now pass it over to Leonard for closing comments. [Audio Gap]

Our top accounts are expanding. Our AI programs are moving consistently from pilot to enterprise scale deployment, and our platform portfolio is deepening both organically and through the capabilities we have added in robotics and Physical AI. I'm confident in the second half of 2026, the strategy is working, the momentum is building, and the team is executing.

Good afternoon, everyone. Welcome to Grid Dynamics First Quarter 2026 Earnings Conference Call. I'm Cary Savas, Director of Branding and Communications.

Joining us on the call today are CEO, Leonard Livschitz; CFO, Anil Doradla; CTO, Eugene Steinberg; and SVP, Global Head of Partnerships and Marketing, Rahul Bindlish.

Following the prepared remarks, we will open the call to your questions. Please note that today's conference call is being recorded.

Before we begin, I'd like to remind everyone that today's discussion will contain forward-looking statements. This includes our business and financial outlook and the answers to some of your questions. Such statements are subject to the risks and uncertainty as described in the company's earnings release and other filings with the SEC. During this call, we will discuss certain non-GAAP measures of our performance. GAAP to non-GAAP financial reconciliations and supplemental financial information are provided in the earnings press release and the 8-K filed with the SEC. You can find all the information I just described in the Investor Relations section of our website.

I now turn the call over to Leonard, our CEO.

Thank you, Cary. Good afternoon, everyone, and thank you for joining us today.

We started 2026 with solid execution, delivering Q1 revenue of $104.1 million that was higher than our guidance range and ahead of market expectations. This performance reflects continued strength in our business model and validates our focus on AI-led transformation and high-value enterprise engagements.

Three trends stood out this quarter, a meaningful and growing contribution from AI revenue, a structural shift in vertical mix toward technology and financial services, and our top customers are undergoing meaningful vendor consolidation with Grid Dynamics emerging as a clear beneficiary.

Last quarter, we called 2026 a pivotal year for the accelerating adoption of our AI offerings. Our first quarter results support that conviction with AI revenue reaching 29.3% of total company revenue, growing nearly 60% year-over-year. Given this concentration and growth trajectory, AI practice has become the core of our business, fundamentally reshaping our offerings, our talent development and our client relationships. I'm confident we are well positioned to further accelerate AI revenues in 2026.

For the first time, our top 5 accounts are entirely outside of retail, reflecting meaningful diversification into technology and financial services, sectors where AI adoption is accelerating and our capabilities are highly differentiated. This group includes 2 leading global technology companies, a global fintech leader, a U.S.-based global bank and a leading financial institution. What makes this group notable is that each of these customers has undergone meaningful vendor consolidation and Grid Dynamics has emerged as a clear beneficiary. This positions us to capture greater market share in 2026 and beyond.

Additionally, we have been actively engaged in AI initiatives across all 5 customers, with some of our largest and most strategic programs driven by this group. Our size and AI technology focus are strategic advantages in a rapidly changing environment. Large enterprises are increasingly seeking highly capable, nimble partners like Grid Dynamics, who can move quickly and deliver meaningful AI outcomes rather than relying on incumbent global system integrators burdened by legacy delivery models.

In many ways, headcount leverage is no longer a competitive moat and differentiation comes from the main knowledge, AI capabilities and ability to rapidly scale relevant expertise.

We're not a systems integrator. We're a product-centric engineering company focused on solving the most complex mission-critical challenges for Fortune 1000 clients with a deliberate emphasis on driving revenue-generating capabilities, not just cost optimization. As enterprises migrate to our custom-developed solutions, the advantage shifts to partners who can build sophisticated production-grade software from concept to deployment. This is precisely what Grid Dynamics does.

AI meaningfully expanding Grid Dynamics addressable market. For example, AI-native SDLC and agentic coding fundamentally changed the economics of delivering services. With delivery time and cost compressing, we can take on larger client initiatives that were previously out of our reach. Also, AI is unlocking a wave of legacy modernization that was not previously economically viable. For years, replacing core legacy infrastructure was considered too expensive, time-consuming and risky. AI lowers these barriers.

At the leading home improvement retailer, the infrastructure for global operations is based on legacy mainframe platforms. Modernizing the legacy mainframe platform was considered risky, and required specialized and expensive talent. Using AI agents, Grid Dynamics delivered a full modernization program within the time line and budget. Grid Dynamics expertise is now extending into physical AI.

In CPG & Manufacturing, enterprises are turning to self-learning robotics and AI technologies to drive operating efficiencies. Our GAIN platform for physical AI makes intelligent robotics more accessible and economically viable. In the first quarter, we closed our first commercial engagement in physical AI with a heavy equipment manufacturer. We're enabling their mining equipment with intelligent autonomous capabilities.

We're building the company around AI. Four pillars define this transformation: AI native delivery, productized engineering, AI consulting, and internal AI automation. The first pillar, AI native delivery, marks a fundamental shift in how we work from human-led workflows to AI agent-driven, spec-based executions across our fixed bid engagements. The economics are compelling and adoption is accelerating. Early indicators point to material productivity gains in select workflows and a structurally different cost base.

In Q1, at our global bank, our autonomous AI workflows analyzed 150 green production applications and uncovered latent defects across systems, including test, and coding and correct behavior. By expanding validated behavior coverage to greater than 70%, we reduced false confidence in system integrity and mitigated production security and regulatory risk.

The second pillar, productized engineering, focused on converting our repeatable IP into AI native platform-based offering under the GAIN platforms. GAIN consists of 4 domain-specific platforms spanning from Agentic AI Commerce, SDLC, Risk and Compliance, and Physical AI. Our engineers increasingly operate as forward deployed specialists composing and customizing these platforms to each client's specific environment, data and workflows. The result is deeper differentiation and stronger client retention.

A good example is that what we achieved in one of the world's largest food distributors. Our client sales associates were spending hours on manual research and proposal preparation for their restaurant clients. We developed AI agents that compressed the preparation process to minutes while improving the quality of the reports. Our efforts resulted in 50% reduction in preparation time and 18% increase in monthly spend for the targeted accounts.

The third pillar is AI consulting. As companies undergo AI transformation, existing business workflows must be evaluated and reimagined for agentic world. Clients are seeking out domain knowledge and deep understanding of AI and data. As a leading global fintech company, our engagement focused on development of AI agents which automate enterprise workflows. Early efforts with our Forward Deployed Engineers embedded inside the client organization have identified inefficiencies and deployed AI agents to automate, optimize and scale the process with a human in the loop, resulting in 15% productivity improvement.

The fourth pillar is tied to adapting AI for our internal operations. Over the past several months, we have been adopting AI tools both off-the-shelf and internally developed in enhancing our productivity and efficiency. This includes areas such as recruitment, RFP responses, knowledge management and HR. With recruitment, we have seen a 2x productivity improvement in terms of number of applicants we can process. With RFPs, we have increased the number of responses by 50% without growing headcount.

With knowledge management, our responses to employee questions improved from hours to minutes. And with HR, multiple initiatives are being rolled out, and we expect more than 20% operational improvement.

Q1 project highlights. Our vertical execution in the first quarter is best illustrated by a few, notable client engagements.

TMT. For a global technology company operating large-scale manufacturing environments, Grid Dynamics designed and validated a unified manufacturing intelligence platform to replace fragmented, manual data flows. The solution is projected to reduce data discovery and reporting cycle times by over 95%. It also lays the foundation for enterprise-wide operational intelligence.

CPG & Manufacturing. Grid Dynamics built and deployed a unified agentic AI platform for a leading global CPG manufacturer, creating the shared infrastructure required to develop, govern and scale AI agents consistently across the enterprise. Running on a major cloud platform, the solution serves as an operational backbone for AI-driven transformation across the manufacturers' supply chain, consumer and commercial domains, the highest complexity, highest impact areas of the business.

Automotive part retailer. For a leading global retailer, Grid Dynamics led the end-to-end modernization of a mission-critical inventory and replenishment platform, migrating from legacy on-premise infrastructure to a cloud-native environment. The program delivered over 70% reduction in infrastructure costs and approximately 40% improvement in core responses time, restoring the platform's ability to support real-time replenishment decisions at the global scale.

At a premier global multi-brand restaurant company, Grid Dynamics deployed an AI coding harness to replace the manual QA workflows that struggle to keep pace with frequent enterprise changes across web and mobile. AI agents continuously simulate customer behavior and adapt automatically to UI modifications in real time, eliminating testing bottlenecks without human intervention. The platform has reduced testing time by approximately 50%.

With that, I will hand over to Rahul Bindlish, Global Head of Partnerships and Marketing, who will share some of the exciting initiatives currently underway and give you a closer look at where Grid Dynamics is headed. Rahul?

Thank you, Leon. Good afternoon, everyone. Partnerships are now a key component of how we go-to-market. Our partner inference revenues have grown to 19.1% of total company revenue in quarter 1, underscoring the value of our ecosystem-driven approach in the agentic era. The majority of our partner inference revenue is driven by Google Cloud, AWS, and Microsoft Azure, our 3 core hyperscaler relationships. They are an active go-to-market channel for our platforms and services. Our go-to-market strategy is aligned with the AI strategy described by Leonard in his comments.

We will be deploying all our platforms on the marketplace of hyperscalers. Our GAIN platform for risk and compliance is now listed on both Google Cloud Marketplace and AWS marketplace. Enterprises searching for production grade capabilities in this domain within those ecosystems will find Grid Dynamics IP directly, increasing our sales pipelines.

We also have joint sales motions with the hyperscalers to accelerate deal closures. That is a fundamentally different way to win business compared to traditional service and sales. This is the first deployment in a deliberate rollout. We are moving additional platforms onto the marketplaces of every major hyperscaler. It also deepens our co-sell relationships with these partners.

Our GAIN platforms plus Forward Deployed Engineers model is a new approach to go-to-market with the hyperscalers. The platform creates the entry point, our engineers deliver the value realization. Enterprises see this clearly and the first few engagement wins reflect their willingness to pay for it.

Each platform we bring to market addresses a specific business pain point with domain-specific IP. This changes the sales dynamics in a way that matters for our growth model. When we lead with a vertical-specific platform, whether that is agentic commerce, compliance or physical AI, we enter a client conversation with a validated solution for a specific business problem. Sales cycles compress, conversion rates improve and initial contracts expand faster because the platform's value is visible to both the business buyer and the technical evaluator. This vertical specificity is what makes our co-sell relationships with Google, AWS and Azure productive.

Grid Dynamics technical depth and domain knowledge, combined with the hyperscalers cloud infrastructure, is what allows us to win engagements against competition. Our AI revenue acceleration is the output of that combination.

We are also expanding our partnership with NVIDIA by porting our solutions onto their software stack. Our GAIN platform for physical AI is built on NVIDIA stack, including Omniverse, and we are taking it to market with NVIDIA for manufacturing and CPG companies.

Industrial AI in manufacturing environments requires simulation fidelity and sensor integration that generic AI infrastructure does not support. Building on NVIDIA's stack positions us to address that requirement and enables joint go-to-market with NVIDIA into a customer segment where the demand for production-grade physical AI is accelerating.

We have also expanded our partnership ecosystem in the AI consulting space, entering into relationships with specialized firms in business process mining and organizational change management. Effective enterprise AI deployment is more than just a technology problem. Clients who deploy agentic workflows are simultaneously reengineering the processes those agents replace and managing the organizational change that follows. By integrating specialized process mining and change management partners into our delivery model, we extend the value that Grid Dynamics offers from platform and engineering, through to adoption and measurable ROI capture.

There are 2 more trends worth noting. Many of the engagements that we are winning through partner channels are extending beyond the initial project. When an AI project delivers clear ROI and our clients are seeing this at scale, the relationship does not close, it expands. Clients return for more use cases, projects and programs. That pattern is visible in our retention data and in the expansion of existing hyperscaler co-sell accounts. At one of the largest food distributors in North America, that pattern played out across 3 distinct phases. The initial engagement was a first project delivered through a co-sell motion with Google Cloud and built on GAIN platform for agentic commerce.

The platform search capabilities were in production within weeks. The client retained Grid Dynamics immediately following go-live to extend the program, using our catalog enrichment solution built on the same platform to improve the quality of the search results. We are now in the third phase, the development of an agentic platform for the client's commercial operations with the first use case targeting sales efficiency already in production.

The margin profile of AI engagements, especially those built on GAIN platforms, is meaningfully different from the traditional services pipeline. When we win through a joint sales motion, clients are buying a validated solution at a fixed commercial structure. That changes the margin profile, higher gross margins than our blended services average. The GAIN platforms plus Forward Deployed Engineers model is not just an acquisition strategy. It's a retention and margin expansion strategy too.

With that, I'll hand it to Anil to walk through the financials.

Thanks, Rahul. Good afternoon, everyone. We recorded the first quarter revenues of $104.1 million, slightly above the higher end of our guidance range of $103 million to $104 million. Our revenues grew 3.7% on a year-over-year basis.

Non-GAAP EBITDA was $12.5 million or 12% of revenues and was at the midpoint of our $12 million to $13 million guidance range. In the first quarter, there was a negative impact from FX fluctuations on a year-over-year basis. We are exposed to a currency basket across Europe, Latin America and India. While we utilize both natural hedges and an active hedging program, the net impact on a year-over-year basis on our EBITDA was a headwind of approximately $1.2 million.

As Leonard highlighted, our top customers are global technology and financial enterprises. And this is by design. Our growth strategy is deliberately focused on verticals where AI adoption is accelerating and our capabilities are highly differentiated. In the first quarter, revenue breakdown reflects this redistribution with meaningful diversification into our TMT and financial verticals.

Looking at the performance of our verticals, TMT became our largest vertical and accounted for 29.5% of total revenues for the quarter with growth of 30.3% on a year-over-year basis. The growth was primarily driven by a combination of our largest technology customers as well as new customers. Retail contributed 28.4% of total revenues in the first quarter of 2026. The finance vertical accounted for 23.5% of total revenues in the quarter, and we witnessed strong demand from our banking and fintech customers. For the remainder of 2026, we are bullish on our outlook with our banking and fintech customers.

Turning to the remaining verticals. CPG & Manufacturing represented 9.4% of quarterly revenues. In the quarter, we witnessed growth from our manufacturing customers in North America and new engagements in Europe. The Other vertical contributed 7.1% of first quarter revenues. And finally, Healthcare and Pharma contributed 2.1% of our revenues for the quarter.

We ended the first quarter with a total headcount of 4,964, up from 4,961 employees in the fourth quarter of 2025 and from 4,926 in the first quarter of 2025. We continue to rationalize our overall headcount as we align our skill sets and geographic mix. At the end of the first quarter of 2026, our total U.S. headcount was 353 or 7.1% of the company's total headcount versus 7.2% in the year ago quarter. Our non-U.S. headcount located in Europe, Americas and India was 4,611 or 92.9%.

In the first quarter, revenues from our top 5 and top 10 customers were 40.8% and 59.7%, respectively, versus 35.6% and 56.6% in the same period a year ago, respectively.

Moving to the income statement. Our GAAP gross profit during the quarter was $36.2 million or 34.8% compared to $36.1 million or 34% in the fourth quarter of 2025 and $37 million or 36.8% in the year ago quarter. On a non-GAAP basis, our gross profit was $36.7 million or 35.3% compared to $36.6 million or 34.5% in the fourth quarter of 2025 and $37.6 million or 37.4% in the year ago quarter. On a year-over-year basis, the decline in the gross margin was from a combination of FX headwinds and higher cost structures across our delivery locations.

Non-GAAP EBITDA during the first quarter that excluded interest income expense, provisions for income taxes, depreciation and amortization, stock-based compensation, restructuring, expenses related to geographic reorganization and transaction and other related costs was $12.5 million or 12% of revenues versus $13.7 million or 12.9% of revenues in the fourth quarter of 2025 and was down from $14.6 million or 14.5% in the year ago quarter. The sequential and year-over-year decline in EBITDA was largely due to a combination of FX headwinds and higher operating costs.

Our GAAP net loss in the first quarter was $1.5 million or a loss of $0.02 per share based on a diluted share count of 84.7 million shares compared to the fourth quarter net income of $0.3 million or breakeven per share based on diluted share count of 86.4 million and net income of $2.9 million or $0.03 per share based on 87.8 million diluted shares in the year ago quarter.

On a non-GAAP basis, in the first quarter, our non-GAAP net income was $7.5 million or $0.09 per share based on 85.9 million diluted shares compared to the fourth quarter non-GAAP net income of $8.7 million or $0.10 per share based on 86.4 million diluted shares and $10 million or $0.11 per share based on 87.8 million diluted shares in the year ago quarter.

On March 31, 2026, our cash and cash equivalents totaled $327.5 million, down from $342.1 million on December 31, 2025.

Since our fourth quarter earnings call, we repurchased approximately 1.8 million shares for a total consideration of $11.5 million. Since our Board authorized the $50 million share repurchase program, we have repurchased approximately 2 million shares for a total of $13.5 million, reflecting our continued confidence in the long-term value of the business.

M&A continues to take priority in our capital allocation strategy. We are committed to augmenting our organic business with acquisitions that strategically enhance our capabilities, geographic presence and industry verticals.

Coming to the second quarter guidance. We expect revenues to be in the range of $106 million to $108 million. We expect our second quarter non-GAAP EBITDA to be in the range of $14 million to $15 million. For Q2 2026, we expect our basic share count to be in the range of 84 million to 85 million and our diluted share count to be in the range of 85 million to 86 million. For the full year 2026, we're maintaining our revenue outlook of $435 million to $465 million.

That concludes my prepared remarks. We're ready to take your questions.

[Operator Instructions] First question comes from Puneet Jain of JPMorgan.

So Leonard, thanks for sharing updates on the GAIN framework. As these platforms become increasingly integrated in your delivery, could you talk about the impact it has on overall operations, say, like are these necessarily fixed price contracts? Do clients pay for tokens like for LLMs or are they bundled in your overall services?

You talked about like Forward Deployed Engineers. Can you train your current employees to be FTEs? Or do you have to change your hiring mix to be able to offer GAIN platform to your customers?

Let me try to unpack some of your questions. It's a lot than one. But let's go backwards, probably a little bit easier. So let's start with engineering talent and Forward Deployed Engineers.

Majority of the people who we deploy, obviously, are internally trained. We have a large number, substantial large number of very technically educated people who we internally build our services and promotions and train them in the models. And it's led by our R&D organization, so you see Eugene is going to give you some more comments, which combining with retraining the delivery organization brings the talent. Obviously, when we bring the talent from the market, it still needs to be structured so they're going to be able to adapt Grid Dynamics GAIN platforms approach.

The GAIN platforms approach is really what makes us different. So rather than talking about a very specific model for each individual customers, let me explain a little bit in the words what these new platforms means for the contracts.

So basically, we developed a lot of tools over time. And even in the last Board meeting, we introduced lots and lots of different names. And now we're maturing to the point that we can offer a suite of solutions to the client where we actually define a kind of a combination of Grid Dynamics IP and open available sources into the total solution. And the total solutions which we offer are driven by adoption of the engineers and agents in the form of the guidance, where we expect the return on investment for the client.

So answering your question, the number of non-T&M projects -- and because there is a lot, there is a tokenization, there is offering of the fixed bid, there is a performance related. They are significantly increased and they continue to increase. And you will actually see that as we continue to answer your questions today because that model itself requires not only training the FD engineers, but adapting the internal processes and the program management and delivery team to actually control a proper engagement in a different venue. So answering your question, definitely, there is a big shift toward non-T&Ms.

The training and rollout of our engineering force is going very successfully. You haven't seen right now from the absolute number of employees, how the dynamics of the headcount has changed yet because number looks flat. But if you again unpack that number, you will see a significantly higher contribution on the engineering workforce because some of them require an additional training and reclassification before we deploy them to the clients. But the good news is, overall, we have a very strong vector where we are building our position with adopting our clients, new models related to the GAIN platforms.

Got it. No, it's a big change. And so it seems like you're already doing a lot of hard work that's involved.

Let me ask Anil. So the guidance, like the full year on top line, so it does imply like a mid single digit growth even in the lower half, mid single digit average sequential growth in second half to hit the lower half of the guidance. So what drives the confidence or the visibility on achievement of this guidance for the full year?

So there are 2 or 3 factors here. Leonard, do you want to talk about pipeline, then I can take it.

Well, I will answer the easy part. And then Anil will dive you a little bit of the numbers. There are 2 parts of the confidence level we have. The number one, the demand has grown substantially. So we have the record number of demand. And I'm avoiding the word number of engineering demand because, again, we're talking about the teams, the platforms, the offering, but overall demand, the vector is very steep right now. That's a subjective factor because, again, this could happen, it may not happen or whatever, but it's a good news. It's a record high.

The more interesting factor is, and Anil will dive into the financial estimates, we are facing a larger, as I mentioned in the previous comment to you, number of non-T&M projects. This work force is defined by a different estimate, how do we qualify the revenue based on this project in which point. So when we unpack the number, we are a bit more conservative, which we're going to guide this particular quarter or the next quarter because now it becomes a little bit more of a financial exercise.

The work has been signed. The work is going on, but Anil probably give you a little bit better feedback. But the summary for you, the takeaway for me, 2 parts, significantly higher number of the pipeline and a very large number of the non-T&M project, which require a little bit more financial attention, how we guide the numbers for the near future for the next couple of months.

No, look, I mean, Leonard, you pretty much hit it. Let me kind of build upon that. Leonard and the team in our prepared remarks talked about a fundamental transformation on how we're moving. And the word you will see again and again is a platform. Now the historical approach we all know is that you take the engineer, you have a certain T&M rate, you multiply it by hours, days; and the formula, as you know, is very linear. We're transitioning. We're seeing that. Rahul is leading the way from a partnership and Eugene is leading the way, obviously, on the CTO.

We've introduced all these new products and platforms, and we're working on monetization. Now there are stages of monetization. There's upfront, that will get start off small. There's greater stickiness with these engineers. And as our clients become comfortable with both our products as well as our engineers in this new model, that's when we start seeing a lot more monetization there.

So when we started looking at these numbers, the obviously, revenue recognition is a key component to it, right? And we're taking, think of it as baby steps right now. We see the pipeline. I look at year-to-date from January 1 through now, compare that with last year, really good. I look at some of these initiatives we're working on, on AI, really good. But the question will be, how do we time it? Is it a linear timing or nonlinear timing? So from that context, for the full year, we're keeping it.

Now let's see the couple of quarters. Does it turn out much stronger because we have some of the recognitions or not. So we're still experimenting with this. We're working through it. So the optics of it looks slightly different from what you can see underneath from a business point of view.

Let me add one more factor, because it could be a bit missed from the first point of view. We also guide substantially better margins. So if you look at the delta between Q1 and Q2, you may ask a question, how can you grow such a steep increase of profitability on relatively modest increase of revenue? So this gives you a little bit more a story that we look at the new projects we've been awarded to us -- as Rahul was mentioning in his statement -- at a different margin profile than the current business. We just don't want to run ahead of the time and do all the financial qualification of that until we see the results. But we are very confident in the progress we're about to make.

So it seems like you are at the cusp of that monetization and that drives the confidence.

The next set of questions comes from Maggie Nolan of William Blair.

I wanted to ask about your partner revenue that crossed 19% of revenue. So where do you anticipate that going? And to what extent do you expect that to be a positive margin driver for the company?

I think the best way to start is with the person who is responding to that. I think, Rahul, you have a perfect opportunity to tell how you build the business continue to grow. So please go ahead.

Yes. Thanks for that question, Maggie. Like you have seen, partnerships have become one of our key go-to-market channels, and it will continue to be. We have a long-term goal to get to about 25% to 30% of our revenues being influenced by partnerships. And we are well on our path to achieve that. In fact, I would say we are tracking slightly ahead when we look at our internal goals to achieve that. And with GAIN platforms being deployed on the hyperscaler marketplaces, we'll probably see acceleration of that partner inference revenues in the future quarters.

Let me just add one more color maybe on this. Rahul, a bit kind of mentioned in his prepared remarks, but it's important because, again, it's new. So we talked with Puneet about the new model of the business. Now we talk a little bit different model of engagement with our partners.

In the past, we've basically been talking about hyperscalers. And that was a very consistent is, frankly, the influence revenue generated with these partnerships. Now we start adding, especially with the physical AI, some interesting new level of partnerships. And monetization is a little bit lower yet, but we see a substantial growth because now we're adding into with the heavy hitters in the industry because it adds more addressable market.

The other element, which is kind of getting also related to our GAIN platforms, it's a consultancy part. So now we're also getting partnerships with some of the business organizations which are asking us to become the lead technology implementation partner, which is adding a little bit more of the flavor from transition from the business conceptual idea to implementation related to specific AI platforms.

As you know, business leaders are a little bit more cautious about spending the budget because you can spend a lot of money on experimentation. So they would like to seek some clarity where they would have a confidence that the investment is not going to be not just risky, but send them to wrong direction. And Grid Dynamics is becoming the partner of that, their consultancy work. So I think it's another really important difference from the past.

On the TMT growth, do you think that's durable into the back half of the year? To what extent was that driven by concentration with particular clients? And what's the visibility into those clients that drove that?

Yes, Maggie, that's clearly a highlight, and it's super exciting. Not only the TMT, but if you look at some of our financial clients there, we have seen many of these customers consolidating. And the other thing is that in some of them, we have now become a preferred vendor. We were always there, but now as they were consolidating, we reached the preferred vendor status.

With the TMT, there are 2 nuances to the movement. There's obviously our work with them, what we're doing. They know what AI is, and they appreciate us. It's a very interesting thing. The smartest technology customers are the one who are seeking our AI capabilities and more, which is a little counterintuitive, right? But the other interesting thing that is going on with these customers is that there's a hyperscaler relationship too. So on both fronts, we are seeing a lot of activity.

Now every quarter, there might be some negatives moving there, but the trajectory is very strong as we get consolidated as we're one of the few vendors, as we've got a clean sheet with many of these new stakeholders and we augment that with some of the hyperscaler growth that is going on.

But I think the important color, very specific color for you, Maggie, is that Anil mentioned about selection being a preferred vendor. We're not talking about generic preferred niche vendor anymore. The AI proliferation equalize the supply base.

In other words, there is -- the size does not provide advantage to some of the largest vendors. The capability of deploying AI solution at scale has been determined as a vital part. And being a smaller company and being able to transition faster remember, again, the very first question from Puneet -- how quickly we can train people. It's amount of quality work with those specialized teams, which determine our awards on the business side. And with the TMT, it's definitely the #1 followed right now with the financial clients. We'll talk a little bit more about others as time comes. But the top 5, top 6 clients, we are in the driver seat for AI deployments.

The next question comes from Surinder Thind of Jefferies.

When we think about the non-time and materials model, how do we think about the incremental risk that you're taking on? Obviously, over the past decade, 2 decades, we moved in that direction because projects got bigger, they got more complex. There is maybe greater uncertainty about scope or changes in scope. How does that work in the new model? Because if you're looking at an outcome-based or fixed price token usage, like where is the risk in the model for you guys? Or how are you guys addressing that?

Surinder, I will actually have Eugene Steinberg, our CTO, to start talking because she is a bit of an architect of the system. And uncertainty has 2 prongs. One of them is a risk level, the second one is a reward level. And I will let Eugene talk about the coexist on both and how we handle it. Please, Eugene.

Yes. Of course, when you are taking a fixed price project, you always have to balance risk versus reward. So on the risk standpoint, the main risks in the fixed price projects are coming from uncertainty. Uncertainty is coming usually from understanding of the requirements and finding gaps in the requirements of the project. We are using very actively our AI agents and our specific game, Rosetta framework, to uncover all the uncertainties in the requirements and clarify with our sources ahead of time during the presale phase, and that builds us a very strong confidence in the understanding of what needs to be done.

During implementation, we are very actively using always AI coding assistance and our GAIN Rosetta framework, helping to accelerate the delivery of a project and building the buffer for any unknown unknowns, which usually happen in those projects.

So let me just add one thing to what Eugene just said. So Surinder, you know you've been in the IT industry, and this is a risk not unique to Grid. It's a universal risk. All I'll add is a couple of additions to what Eugene said. The first thing is that when you scope out projects, if you don't have a deep understanding of the project or as Eugene says, the risk, it's a problem.

Now when I look back at the history over the last 5 years, historically, we were a T&M shop. We moved towards fixed price. And actually, during those first year or 2 of our fixed price, we learned a lot. We have committed mistakes in the past. This is the pre-AI era, and we worked. As a matter of fact, there were times when our fixed price project margins were comparable with our T&M, and I always went back to the team what's going on. So we learned. Now when you look at our fixed price margins pre-AI, they're higher than our T&M. And those learnings are now moving into our AI.

So we really know what we're doing. I think what we've learned is that if you don't understand the problem that you're dealing with and you don't have a technological know-how, you're absolutely right, there is a heightened level of risk. We'll always have that risk. But as Leonard pointed out, there's a reward component too with that.

Yes. And I just want to close on that with one simple statement. In my prepared remarks, I mentioned clearly that Grid Dynamics is not a system integrator. We are a product-centric engineering company. And that actually gives us the higher level of confidence that we take on the projects, we have a higher probability of success.

So Eugene was mentioning Rosetta, another methodology we're using. It's all part of the GAIN platforms. Now the outcomes on a greater scale, Surinder, will be seen as we will propagate more and more results of this work. So it's not about how much money we generate in the project, but how much rate of growth we're going to see in this project going forward. Right now, at the size that we have and the scale of the tasks, we are training not only the models, but our customers, how to react on gradual, I would say, continuation of the development and approaching the goals.

So it's very, very important for the fixed bid for us to make sure we have intermediary goals because the approximation of the work and deliver results have to be iterative process. And that's very important. So we're improving not only our technology capability, but our project management relationship with the clients as well.

Maybe just a quick related follow-on. Any color or commentary on the delta between kind of the fixed price margins that you're able to achieve currently and what you're achieving on the time and materials side?

Sure. So when I look at -- now it varies quite a bit, right? So I'll throw a number out and somewhere in the ZIP code. I have seen the contribution margins when we get to some of our AI work somewhere in the 60-plus range too. Now I mean, not every project is a 60%. Otherwise, we would have been a 60% gross margin, but this is a contribution margin and then obviously, you have to offset by some of the overhead.

In general, if you look at most of our AI work, it is higher margins. If you look at the deltas between our T&M business and non-T&M business, there is a delta. So we see non-T&M in general being higher. And then when you look at AI business portions of the business, we do see some outliers, very positive outliers.

Ultimately, what does this mean from a gross margin perspective? There's obviously the near term that you're able to handle from both managing headcount. But can you talk about where utilization is relative to your headcount goals and how we should think about the evolution over not just next quarter, but the next 12 to 24 months? Because it sounds like there's a big opportunity here, and I just want to make sure I understand the component that you control through managing headcount and utilization versus the component that's ultimately going to roll out as a result of just the revenue mix itself.

Very good question. So the way I look at, Surinder, your question is there is what I call the near to intermediate areas of focus, which is part of our 300 bps margin expansion, right, Q4 to Q4, and you're already seeing that, right?

Then there's a more fundamental question that you're asking is what is this pricing model and what is the margin model. So that is a more evolutionary thing that will not happen overnight, that has a more longer term. And that is what we are all working on as we work on these AI platforms.

The whole GAIN -- as a finance guy, if you really look at what I tell Rahul from a GAIN platform and Eugene, who's always excited about technology is, what does it do to the margins and what does it do to the stickiness and what does it do to the growth? I mean, that's what it really boils down to, right? And our long-term model is to embed GAIN platforms with our customers -- that is just not human capital, but it's agents and actually IP -- create more stickiness, move towards a more fixed price model, which should result in a higher margin structure. Now what is that finally going to end up being? It's work in progress.

Yes. So I think Anil gave you a lot of financial guidance. Let me break it down to a couple of key elements, which I gauge the business.

So there are 3 elements, obviously, adoption of AI in terms of the efficiency of the business, the marginality of the business. But there's a third factor, which you guys use quite often, which is not totally irrelevant. I think it's quite appropriate. It's the revenue per person. So utilization of the test becomes more driven by the revenue per person increase. And there are 2 parts of it.

On an overall EBITDA margin on a net margin, this is the fourth pillar of the platform, how internally we utilize it. But that doesn't help with the growth of the business. With the growth of the business, it comes actually with the idea that we are going to have repeatable and kind of reusable IP intelligence of our platforms. So the utilization part comes with the utilization of humans and IP capital. So it's a new formula, which is really -- will be gauged in my opinion, which I'm going to drive the company -- is increased revenue per person.

Now saying that, there's another factor, right? It's Europe versus India versus U.S. local consultancy. Different categories of different regions create a different ratio between revenue and the margin. And I'm telling my team, it's irrelevant. The revenue per person as a guidance for utilization has to grow everywhere. The new ability to create game-based platforms Forward Deployed Engineers and the models should drive the efficiency as we already see in the early adoption regardless of the regions and the traditional T&M models, which are not going to be as much used as we go forward.

The next set of questions comes from Bryan Bergin of TD Cowen.

Maybe just at a high level to start on client sentiment. Just given the war in Iran, anything you can comment on how the conversation with enterprises has progressed over the last 2 months here? And just more recently as well, anything in recent weeks that's different?

Yes, I can do that. Thanks for that question, Bryan. So there are clear trends, Bryan, that we are seeing with our clients. Number one is whereas last year, there was clearly clients who were looking at AI projects as POCs and trying to progress them into projects. Clearly, this year, there are production projects being invested in clients across the industries, very consistent.

Second trend we are seeing is with AI, it is driving more projects and programs even for application modernization and data platforms. So we are seeing our pipeline grow in those 2 areas as well.

Third, very clearly we are saying -- whereas the last year, they were the early adopters of AI, now we are seeing a wave of fast followers. That is increasing really our pipeline as well as, in some ways, our total addressable market.

Bryan, coming to your point, the Iran war, to me, at least when I look at the business, it's a non-event at this stage, right, in the third place.

Yes, I would say I would not really comment right now because the situation is very fluid there. We don't conduct the business in an area of the direct impact. So it's very hard to say that. The secondary impact on the business, again, it's negligible. I think that we had a huge impact continuing to the impact of the Russian invasion to Ukraine, right? That's much more dear to us. I don't think we're affected as much. But the global world has changed more with the conflict of Middle East and obviously conflict between Russia and Ukraine. And there are various factors.

I mean, look, ultimately, the peace and resolution is the benefit for everyone. But how the peace is going to be achieved is very important. Right now, we're just plugging alone. And in our business model and our customer relationship, there is no detriment. There are some positive movements related to their retooling, especially in the manufacturing space because there are obviously more demand for manufacturing of certain type of products.

If we talk about our digital twin approach and about our physical AI approach, we're gaining momentum. But I would hate to say that it's really driven specifically by the individual event. But we definitely see the shift of manufacturing to the much higher retooling and scaling the production. And one of them is related to the traditional manufacturing. One of them is related to more semiconductor manufacturing.

Second question here, just as it relates to kind of the AI productivity conversation, just coming out of a lot of the larger traditional SIs, the conversation around productivity, pricing compression for them became more pronounced here in recent weeks. I fully understanding you're not competing in many of the places that they are. But just how are the enterprise conversations for you in engagements that are not transitioning under the game framework as far as that type of a dynamic?

So how the conversations are going in the framework -- so in this case, very often, we still enjoy significant productivity improvements from AI. I can give you some examples. So we just completed a project with one of the wealth management client of ours. And this is where we deployed AI agent across the CA pipelines in one of their large business units. So there, we saw 3x to 6x productivity improvements in the creation of the test coverage. And that allowed us to go wide in this customer and increase our stickiness and increase our reach to all business units of these customers going forward. That proved that we can do more with less resources and this differentiates us across other vendor base of this customer.

Yes. So let me add a couple of statements to what Eugene just said. So the question is really how is the pricing environment right now beyond the AI. So AI obviously has its own dynamics, and I will put that aside.

When I look at the business, I look at a couple of very interesting things. One is that I do not see clients coming and asking that now that same engineer give me a big discount now. I'm not seeing that. Now we can argue whether I'm seeing a premium or more premium, that's second question. But we're not seeing any pricing pressures.

Number two is that in our case, tied to Leonard's opening comments, we've seen a lot of vendor consolidation over the last 18 months. Very interesting thing about vendor consolidation, it's good news and not so good news. The good news is that they go from hundreds to dozens. The bad news is that, okay, they say that you're one of the chosen one, give me a little bit of a discount for the next year or so, something like that, right? So we've gone through that.

So I would say maybe that would be the closest thing I could come to. But the team does a very good job when it comes to new customers, new logos. They're very particular. We have a very strong discipline in terms of ensuring that the margins come in. It's with our well-established customers. And there, we're seeing some of these trends.

You have a very clear example now.

Yes. I just want to add a couple of points there, Bryan. Number one, productivity improvement in the industry is still being shown at individual developer level. When you translate that into projects, especially brownfield projects where majority of our business is, where you are integrating into legacy systems, that productivity at a project level actually falls down to significantly lower numbers, right? So from that perspective, there is less pressure because you are executing projects and programs and not providing individual engineers.

At the same time, when we have examples of consistently showing productivity improvements, we are able to go back to our customers and grab more business. So it becomes expansion of a business strategy rather than play on the margin or the rate.

I think let me just conclude. In a good environment people talk about their side cases and I kind of summarize from the global business positioning. So what I see, and this is quite promising because when I personally meet with the leaders or clients and usually, when you go to the top, the conversations on the overall spendings, and the priorities and budgets come quite clearly as a critical path, especially when those leaders coming from technology organizations, which depend to show concrete results to their business leaders. They are much more focused on productivity in terms of the overall return to the clients.

Remember, we talked about this in the past. So you agree with business people on ROI on a total budget versus outcome and then you go to the VMO, and VMO breaks it down by the rate per person. We are getting right now in a budget discussion overall projects, where the budgets are driven by the fixed bid by the deliverables. And that model, that productivity conversation usually goes on a deployment of the measurable results before somebody starts looking at productivity, because when are you going to ask productivity if it's a total budget being agreed between both sides.

So this environment a little bit better. But before when Surinder was talking about, he acknowledged, obviously, the question of the risk of the model. But that risk is not related directly to productivity anymore at those new adapted businesses.

I've got one last one for Rahul here since he's on the call. Just Rahul beyond the major hyperscalers, as you think ahead, what other types of partner ecosystems are you focused on?

So I think there are going to be at least 3 categories. I already spoke about NVIDIA. I do expect that partnership to take off from here.

The second category would be specialized partners. I talked about on the AI consulting area. But I do expect as technology evolves, there are more specialized AI firms that we will start to partner with, potentially even the likes of your LLM providers, right, as their strategies evolve.

The third category is what Leonard had talked about. We are starting to see interest from large consulting business consulting companies who are looking for technology partners to enable capabilities that they want their clients to have, right? And that's the third very interesting partnership area that I see us progressing with.

This is immediate. This is we're developing right now.

This is we're developing right now, yes.

The next questions come from Mayank Tandon of Needham.

I don't know if there's much to ask. But I'll go ahead anyway, give it a shot.

Mayank, we expect you to be the best questions.

I'm sorry, I'm running out of questions here. But I guess just very quickly, just to keep the call on schedule. The question I had was around your visibility. I think you talked about that earlier, Anil.

In terms of the revenue, how much of the business would you say is sold versus you have to still go out and win? So what is sort of potentially at risk versus what you already have in the bag in terms of your guidance?

So you recall, Mayank, we have had a very traditional model or a well-established model about 85%, 10% and 5%, right, where 85% of our revenue in any given year comes from customers who have been with us 2 years and beyond, 10% comes from over the last 12 months and 5% comes from new. That framework more or less continues to be intact. There might be some variations, especially as we ramp some of these new customers. So the way -- I look at it through this lens.

Now when you look at our whole guidance philosophy and when you look at our whole outlook philosophy, what we know well is potentially where we have some of these downside risks, right? I mean, we're dealing with these customers and these are big customers, and we have some sense of what we do.

So when we give our guidance, for example, at least in the short term, we're taking that into account. When I switch from my short-term guidance to my long-term guidance, I basically switch from a bottoms up to a top down a little bit, right, where I look at the overall pipeline, I look at the forecast, I look at our customer engagements and come up with this.

Now if you were to ask me whether I have a number that I believe is at risk, I mean, it's a whole probabilistic distribution, right, on how I look at it. I would say when I look at the business today versus 3 months ago versus 4 months ago, things are improving. So qualitatively, I would say that things are improving.

Now there's always that risk that we have with any one particular customer due to circumstances or as someone asked a question on the Iran war, there's a macro issue, consumer-sensitive industries are impacted. That's always there. But as we see right now, we feel good about where we see the overall business.

So let me just give you, as always, direct pointers. After listening to Anil we need some guidance on his guidance. There are 2 areas which I think are very important to understand. Number one, the retail business, which traditionally was the most volatile has been derisked and continues to be derisking because it's a smaller contribution. It's not little, but it's small. So that's area where the variance of uncertainty you are talking about. But the second risk is actually growing as we're going to grow the business is how the AI deployments will actually convert into the measurable profits and gain, not Grid Dynamics GAIN platform, but the client gain, right? And that business is growing very fast.

So we're very happy that we can actually forecast a better deployment of these projects. But again, when we talk about fixed bids, we're talking about outcome-based, we're talking about criterion, which before was not that clear, exactly it's how do you measure that ROI. So this criterion becomes a system of criteria, which is growing more and more of our business.

So I would say that the business we project is very certain that we're substantially derisking with retail. However, I see as we grow macro going forward, we need to make sure we bet on the right partners. And that's when actually the ecosystem of the partners also evolves. Remember, Bryan's question, who is going to be the next level of partners besides hyperscalers. And then Rahul mentioned 2 parts, of course, consulting is very clear gain. But then which of the other elements of the LLMs on other substantial guys who will provide us data centers, who provide us the material traffic of these deployments, the cost of these models is going to play a much bigger role.

We are tuned to the system. We're selected to be preferred in many cases. We're confident. But the whole dynamics of AI deployed deliverable value, it's still something we have to prove on a major scale for everyone.

Just to close out, Anil, you mentioned that M&A is still a priority for you. So just wanted to get some context in terms of what you might be looking for. And then, have private companies maybe sort of recognize that valuations have come down a lot and maybe are more inclined to sell versus resisting a potential sale to a company like Grid?

Yes. So as you rightly pointed out, yes, we're very focused, fingers crossed. We hope to close some deals -- and most of them are tuck-ins. What we're looking at right now are tuck-ins from a capability point of view. So obviously, technology has elevated to be very important, data, AI and certain end markets tied to our strategy.

So now when it comes to the valuation, you will always have to pay a premium for good companies. For good, capable companies, you will always have to pay some level of premium. But overall, you're right, they have come in. And things are looking better from a valuation point of view. But at the end of the day, if someone has some true differentiation, you do have to pay.

The bottom line is, the accretiveness of these acquisitions have been the vital point, and we're very close to prove to the market we can still come back and do our M&As because, again, you're right, the appetite for them has been a little bit more modest, but it's not as critical as our broader net, which we threw around the world related to the 2 elements, really 2 elements: AI-related technologies, especially the cutting-edge technologies, we can benefit more as a congruent business than the particular company on themselves. And the second part is looking for the partnership outside of the traditional path, which we're enhancing. So stay tuned. We're in good shape with that.

Ladies and gentlemen, this concludes the Q&A portion of our call. I will now turn it over to Leonard for closing [Technical Difficulty].

Q1 2026 is proof that our AI transformation is working. Our revenue reached 29.3% of total revenue. GAIN has matured from a framework to platforms with Forward Deployed Engineers. Agentic AI solutions are now in production across a range of industry verticals and are generating measurable ROI at commercial scale.

The pipeline entering Q2 is the strongest it has ever been. AI consulting and hyperscale partnerships are expanding. We're executing on our strategic road map, including AI-native delivery, productized GAIN platforms, consulting and internal automation.

We look forward to updating you next quarter. Thank you.

Good afternoon, everyone. Welcome to Grid Dynamics Fourth Quarter 2025 Earnings Conference Call. I'm Cary Savas, Director of Branding and Communications. [Operator Instructions]. Joining us on the call today are CEO, Leonard Livschitz; CFO, Anil Doradla; CTO, Eugene Steinberg, COO, Yury Gryzlov; and SVP, Head of Americas Vasily Sizov. Following the prepared remarks, we will open the call to your questions. Please note that today's conference call is being recorded.

Before we begin, I would like to remind everyone that today's discussion will contain forward-looking statements. This includes our business and financial outlook and the answers to some of your questions. Such statements are subject to the risks and uncertainty as described in the company's earnings release and other filings with the SEC.

During this call, we will discuss certain non-GAAP measures of our performance. to non-GAAP financial reconciliations and supplemental financial information are provided in the earnings press release and the 8-K filed with the SEC. You can find all the information I just described in the Investor Relations section of our website. I'll now turn the call over to Leonard, our CEO.

Thank you, Cary. Good afternoon, everyone, and thank you for joining us today. I'm delighted to share that Grid Dynamics closed 2025 with another landmark performance. In the fourth quarter, we beat Wall Street expectations on both revenue and EBITDA delivering record revenue of $106.2 million and a strong $13.7 million in non-GAAP EBITDA.

Remarkably, we finished the full year with a record revenue of $411.8 million, which is 17.5% growth year-over-year. 2025 non-GAAP EBITDA was $53.8 million. In Q4, our top 3 customers, including two global technology companies and the largest payment technology company. All of them are leaders in the AI space. Our performance is a result of our AI expertise, the strength of our accelerators and kind domain knowledge.

In Q4, our AI revenue grew 9% over Q3 and now represents 25% of our overall revenue. For the full 2025, our revenue reached over $90 million, representing 30% year-over-year growth. In 2026, when dissipate continued AI revenue growth.

There are three key factors driving our bullish outlook on AI. First, AI coding agents and automation, significant enterprises build versus buy calculus to build at a lower cost. The shift aligns with Grid Dynamics core strengths in building solutions for Fortune 1000 companies, leveraging specialized talent and intellectual property.

Second, our efforts with [ Gain ] are resulting in a richer blend of outcome and output-based engagements. Crucially, these new engagements enable us to decouple pricing from effort. We have successfully deployed software platforms across multiple industry verticals. Our AI engagements now strategically combine the strength of our human capital with the value of our platform assets.

The market reception for these software platforms has been strong, with customer demonstrating a clear willingness to pay. This positions us well to grow recurring revenue, deepen customer retention and extend the duration of our engagements. Grid Dynamics engagement structure will contribute to our 2026 margin expansion.

Third, the speed of AI transformation is not uniform across industry verticals. While we continue to generate revenue from the retail and CPG verticals, we prioritize investments in the area of technology, financial services and manufacturing, where we see significant opportunities for customized auditable product-grade agentic AI platform.

Let me talk about Grid Dynamics vertical strengths. In the price we're learning that deploying AI at scale requires deep domain expertise. We cannot build an effective Agentic system for a production floor without understanding manufacturer. You cannot build one for a global permit network without understanding the compliance architecture. Such expertise is what we have been building vertical by vertical for nearly 2 decades. Now we're qualifying it into platforms.

Our Merchandising Experience Platform, MXP, brings our expertise to marketplaces and digital commerce. While XTDP, [ orbitemporal ] Data Platform help financial clients, specifically in capital markets with auditability and other compliance challenge. Platforms unlock IP-driven outcome-based engagements, and that's how Grid Dynamics move from billing for effort to billing for value.

Now let's talk about partnerships. Our partner influence revenue reached a significant milestone in 2025, exceeding 19% of our total revenue. So significant growth underscores our mission to keep Grid Dynamics at the forefront of modern enterprise infrastructure. We have strengthened our relationship with all hyperscalers through targeted investments in Agentic platform capabilities, earning specialized badges and building new joint solutions.

Notably, in December, we signed a strategic collaboration agreement with AWS for data foundations in AI. Our premier partnership enables Grid Dynamics to receive funding from AWS to support AI enterprise initiatives.

Our collaboration with NVIDIA on Omniverse based solutions is enabling us to deliver high fidelity industrial-grade digital twins that are essential for our physical AI expansion.

In the fourth quarter, our vertical execution is best illustrated by several notable projects. Fintech transformation. We partner with a global financial leader to launch a proprietary generative AI agent supporting more than 10,000 financial advisers. This interactive experience replaces static policies with personalized guidance as is projected to increase productivity by about 20%.

TMT Analytics for a global technology enterprise, we modernize a legacy mobility application into a scalable analytics platform. Providing centralized visibility into global travel activity and spend. The platform has materially improved usability, increased feature velocity and reduce stakeholder coordination overhead.

Discrete management. We developed a comprehensive dispute management solution for a leading financial services firm. By integrating Generative AI, the platform streamlines charge-back challenges, increasing win rate and reducing operational overhead. Financial governance. At a leading U.S.-based global bank, we're building a global agent runtime and AI orchestration platform, enabled business-focused agent to automate complex workforce starting with successful automation in internal compliance. We also deployed the age-driven executive insight capabilities that provide leadership with consolidated global operational summaries.

With that, let me turn the call over to Eugene Steinberg, our CTO, who will talk about our AI capabilities, how we are upskilling our engineering workforce and how we're using it to improve our internal operations. Eugene?

Thank you, Leonard. Good afternoon, everyone. We are actively executing across three horizons. AI first engineering, Agentic Enterprise and Physical AI. In Q4, we shipped across all three, and these foundations position our AI business for 2026.

Horizon 1, AI first engineering. Horizon 1 is the core of our current business. The engineering work that source the majority of our clients today. We are accelerating productivity across the organization through AI first native tooling and investing decisively in the continuous upskilling of our engineers.

Enterprises are no longer debating the merits of adopting AI for software development. But rather how to do it without losing control of quality, security and institutional knowledge.

It is in this context that we launched Rosetta our AI native software development framework. Rosetta is part of our [ Gain ] initiative and provides a governance layer for AI coding agents. Rosetta automates contract setup, enforces consistent workflows and manage his engineering knowledge at both the engineering and organization level. It operates within the client's own security perimeter and works across all major coding platforms.

Developers get consistent project aware agent behavior from day 1. Engineering colleagues get centralized governance and visibility across the entire agent footprint. With Rosetta clients benefit from decades of institutional expertise seamlessly embedded in the way engineering workflows. We have several engagements underway and a scaling gain as the standard delivery backbone across all engagements in 2026.

Grid Dynamics separations is client 0 for our AI solutions. Cerebra, our internally developed Agent platform launched in Q3. It is built on Google AI stack, Gemini enterprise, ADK and A2A. Within Grid Dynamics, Cerebra is being used by our sales recruitment and knowledge management organizations, automating proposal development, technical prescreening and research at scale. Clients adopt faster than the platform has already been stress tested in production. As AI revenues ramp, we expect this model to drive both revenue growth and margin expansion.

Horizon 2, Agentic Enterprise. Horizon 2 is where we are expanding and investing by leveraging our engineering debt to enterprise transformation at scale. The Agentic area is reshaping the economics of software delivery. AI native development tools are allowing the overall cost of building and deploying software, placing pressure on systems integration and configuration programs.

At the same time, client expectations are rising. Programs previously too expensive or too slow to justify are becoming feasible. Enterprises are thinking bigger and moving faster taken on significantly larger mandates. That means moving away from SI-heavy engagements and toward a regional in-house engineering.

That rotation plays directly to our strength. In the past decade, enterprises have increasingly became dependent on system integration, assembling Software-as-a-Service ecosystems, configuring cloud services, and stitching together vendors products. In the Agentic era, this changes fundamentally. Production deployments require bespoke engineering. Purpose-built agent workflows the main specific data and knowledge layers, distributed system and platform engineering.

Grid Dynamics is well known for its engineering capabilities and proprietary IP at leading global enterprises. The genic area rewards builders, and that is where we have invested.

Our go-to-market runs two tracks. For Tier 1 enterprise clients, we architect and could develop custom verticalized AI platforms built around the specific architecture, governance and compliance requirements. For Tier 2 mid-market clients, we integrate hyperscaler platforms with Grid Dynamics verticalized components on top, optimizing time to value and overall cost.

Both tracks are expanding. We have also established a partnership with temporal through the JumpStart program. This initiative positions us as a technology consultants for [ Tempur's ] customers. embedded in crucial architectural decisions from the outset.

This partnership has generated multiple new engagements across financial services, enterprise software and industrial sectors. The proof points are concrete. A notable example is our work with one of the world's largest payment networks, where we are leading a broad Agentic AI program. We have developed a required service across 17 applications, a universal enterprise assistant with agent to agent communication and centralized governance and evaluation.

Our efforts have led to an approximate 40% reduction in build time and 60% reduction in ongoing maintenance efforts. [ Visa ] platform deployed across 30,000 employees. The impact has been measurable. Specialized groups are seeing up to 15% productivity improvement, driven by faster information access and reduced manual research.

As a leading global CPG company, we developed over 20 enterprise-ready AI agents through a unified agent factory platform. This delivered 15% productivity improvement across enterprise users. These deployments confirm a pattern we see consistently. Once AI capabilities more fully in production, clients realize approximately 15% productivity gains, tangible operating leverage at enterprise scale.

We are leveraging our deep domain expertise to build vertical air platforms, co-defining patterns in the structured productized offerings. Our initial solutions have real traction and are generating revenue with enterprise clients.

MXP, our merchandising and product discovery platform illustrates its progression most clearly. It began as search engineering expertise, evolved into reusable accelerator. And in 2025, gross intel license revenue, this is a growing customer base across North America, Europe and Latin America. Its deployment for a leading European luxury retailer delivered a 7% total revenue uplift, and a 50% reduction in merchandising workload, while handling a 25% year-over-year surge in peak holiday traffic without disruption.

[ XTDB ] is our platform designed for the financial industry, a bitemporal database built specifically for regulated financial environments. As financial institutions deploy AI agents, the regulators require full point and time reconstruction of any decision. Banks deploying agents for trade processing, compliance or investigations, need systems that can capture precise information related to trading activities. [ XTDB ] addresses that with full auditability across both business time and system time. The platform has been adopted in several global banks and in Q4, we shipped a significant new version extending its capabilities for multi-entity data mesh environments.

It is this kind of deep infrastructure IP that differentiates our financial services practice from generic AI Consulting. Our engineers no longer arrive as individual contributors. The if backed by codified IP, Rosetta, MXP, [ XTDB ] and documented patterns from dozens of deployments. The client gets immediate expert deployment, not a learning curve.

Horizon 3, Physical AI. Horizon 3 is our forward-looking investment in Physical AI. Bringing the same AI engineering depth they apply in software to industrial and manufacturing environment. Our flagship platform here is Incarna, a software platform that supports the robotics industry. Incarna dramatically compresses with time required to program robots for complex manufacturing tasks, enabling robots to handle high variability physically demanding work that conventional automation cannot address.

In partnership with Smart Ray, a leader in industrial 3D vision sensors we developed and deployed the Incarna AI model for robotic weld inspection. Weld inspection is demanding. Commodity requirements are stringent and variability in materials and geometry makes rule-based automation unreliable. The result, high inspection consistency, improved quality assurance and scalable automation in environments where precision is nonnegotiable.

At the Fortune 10 manufacturer, we automated the conversion of CAD files to see in [ C ] machine instructions, a workflow that previously took 5 days now completes in hours, greater than 90% cycle time reduction, validated in production. We will have more to share as this program scale.

As we look ahead, we will build on our foundations. We are rapidly and deliberately scaling towards a multi-industry LED business transformation. [ Gain ] in [indiscernible] our engineering judgment, so its skills beyond individual engineers. MXP shows that our IP can generate revenue as software, not just as a services. [ XTDB ] gives us a technically differentiated entry into finance. Incarna, opens doors and manufacturing. And our Agentic practice is shifting from the spoke delivery to structured vertical offerings where our accelerators compress time to value and our contracts increasingly capture outcomes. We are moving from labor scale growth to IP scale growth, and that transition defines our 2026 execution.

With that, let me turn over to Anil.

Thanks, Eugene. Good afternoon, everyone. We recorded fourth quarter revenues of $106.2 million, slightly above the midpoint of our guidance range of $105 million to $107 million. This represents a sequential growth rate of 1.9% and a year-over-year growth rate of 5.9%. There were 30 bps and 22 bps of FX headwinds on a sequential and year-over-year basis, respectively.

Non-GAAP EBITDA was $13.7 million or 12.9% of revenue and was at the higher end of our $13 million to $14 million guidance range. In the fourth quarter, there was a negative impact from FX fluctuations on a year-over-year basis. We are exposed to a currency basket across Europe, Latin America and India.

While we utilize both natural hedges and an active hedging program, the net year-over-year impact on our EBITDA was a headwind of approximately $1.5 million.

On a sequential basis, there was a tailwind of approximately $160,000 to our EBITDA as the dollar strengthened relative to the British pound and euro. Looking at performance of our verticals. Retail remained our largest vertical, contributing 28.7% of total revenues in the fourth quarter of 2025. While revenues in this vertical increased by 5.3% on a sequential basis, there was a decline of 6.9% on a year-over-year basis. The sequential increase was broad-based across our retail customer base.

TMT, our second largest vertical accounted for 28.3% of total revenues for the quarter. The vertical delivered strong results with growth of 5.3% on a sequential basis and a 27.5% increase on a year-over-year basis. The strong year-over-year growth was primarily driven by our top 2 technology customers.

The finance vertical accounted for 22.9% of total revenues in the quarter, growing 5% on a year-over-year basis. This growth was primarily driven by increased demand from our large fintech customer and large banks.

Turning to the remaining verticals. CPG and manufacturing represented 10.2% of our fourth quarter revenues. This vertical remains stable in absolute dollars sequentially but declined 4.3% on a year-over-year basis. The year-over-year decline was largely due to a decline at some of our automotive customers. And this was partially offset by our CPG customers. The other vertical contributed 7.3% of fourth quarter revenues. This remained flat on a dollar basis relative to the third quarter and grew by 8.4% on a year-over-year basis.

The year-over-year growth was primarily from our meal kit client. And finally, health care and pharma contributed to 2.6% of our fourth quarter revenues. We ended the fourth quarter with a total head count of 4,961 slightly down from 4,971 employees in the third quarter of 2025 and that from 4,730 in the fourth quarter of 2024. Although our total head count was down on a sequential basis, our billable head count increased meaningfully. We continue to rationalize our overall head count as we align our skill sets and geographic mix.

At the end of the fourth quarter of 2025, our total U.S. head count was 357 or 7.2% of the company's total head count versus 7.4% in the year ago quarter. Our non-U.S. head count located in Europe, Americas and India was 4,604 or 92.8%. In the fourth quarter, revenues from our top 5 and top 10 customers were 39.7% and 58.5%, respectively, versus 35.6% and 55.8% in the same period a year ago, respectively.

Moving to the income statement. Our GAAP gross profit during the quarter was $36.1 million or 34% compared to $34.7 million or 33.3% in the third quarter of 2025 and $37 million or 36.9% in the year ago quarter. On a non-GAAP basis, our gross profit was $36.6 million or 34.5% compared to $35.2 million or 33.8% in the third quarter of 2025 and $37.6 million or 37.5% in the year-ago quarter.

On a year-over-year basis, the decline in gross margin was from a combination of FX headwinds and greater mix of U.K.-based head count from our acquisition of JUXT.

Non-GAAP EBITDA during the fourth quarter that excluded interest income expense, provision for income taxes, depreciation and amortization, stock-based compensation, restructuring expenses related to geographic reorganization and transaction and other related costs was $13.7 million or 12.9% of revenues versus $12.7 million or 12.2% of revenues in the third quarter of 2025 and was down from $15.6 million or 15.6% in the year ago quarter.

The sequential increase in EBITDA margin was from a combination of higher gross margins and FX tailwinds. On a year-over-year basis, the decline in EBITDA margins was largely due to a combination of lower gross margins and FX headwinds.

Our GAAP net income in the fourth quarter was $0.3 million or breakeven per share based on a diluted share count of 86.4 million shares compared to the third quarter net income of $1.2 million or $0.01 per share based on a diluted share count of $85.8 million and net income of $4.5 million or $0.05 per share based on 83.8 million diluted shares in the year ago quarter.

On a non-GAAP basis, in the fourth quarter, our non-GAAP net income was $8.7 million or $0.10 per share based on 86.4 million diluted shares compared to the third quarter non-GAAP net income of $8.2 million or $0.09 per share based on 85.8 million diluted shares and $10.3 million or $0.12 per share based on 83.8 million diluted shares in the year ago quarter.

On December 31, 2025, our cash and cash equivalents totaled $341.1 million, up from $338.6 million on September 30, 2025. M&A continues to take priority in our capital allocation strategy. We're committed to augmenting our organic business with acquisitions that strategically enhanced our capabilities, geographic presence and industry verticals.

Coming to the first quarter guidance, we expect revenues to be in the range of $103 million to $104 million. We expect our first quarter non-GAAP EBITDA to be in the range of $12 million to $13 million. For the first quarter of 2026, we expect our basic share count to be in the range of 85 million to 86 million and our diluted share count to be in the range of 87 million to 88 million.

For the full year 2026, we are bullish in our outlook. We expect revenues to be in the range of $435 million to $465 million.

That concludes my prepared remarks. We're now ready to take questions.

[Operator Instructions] The first question comes from Maggie Nolan of William Blair.

So you've had impressive growth in AI revenue and you're above $90 million for 2025. So I'm wondering if projects are moving into production at scale and then what is the nature of these projects? And how is the demand among customers?

Look, we extensively discussed in various forms what AI represents to green dynamics and what is the opportunity for us going forward. Fundamentally, what makes a big difference for Grid Dynamics for 2026 on is that we're not only moving from the small development project to full scale implementation, but also we introduced our platforms, which has been noted during this particular time. And that kind of scales the confidence with the clients to give us more of the solutions where we represent our engineers combined with their own tools as a new way to building the solution faster and more affordable for the clients. Perhaps some words from Eugene.

Yes. It's a great question. And there are two main zones, which are most exciting for me. One is AI-powered customer experience. The reason behind that is that this is the zone where the impact from source personalization, Agentic commerce is very obvious and memorable by our clients. And this is where clients see ROIs in weeks, not in months or years. And that allows us to expand those accounts very, very quickly based on this successes which we see in this domain.

Second is enterprise platforms, not as visible as front end work or our customer experiences. But this is a foundational layer, which helps our companies to organize their data, build AI agent factories on top of this data and then go into developing business agents on top of those platforms. And what we see in our projects is as we taper mature and go to production clients start to scale very, very quickly building the agents, and we are helping them to develop as AI agents. And we are going from 1 to 10 to 20 of those specific customer facing, which will collect agents very, very quickly. So this expands our work and allows us to move very, very quickly.

Great. And then anything else you would comment on as you move into 2026, kind of how you expect the trend to evolve any way that you can maybe tie that back to the numbers or maybe some of your margin expansion goals you've mentioned.

Yes. So we marked you, with bunch of names during this press release, right? So we were talking about merchandise is experienced platform, we're talk about [indiscernible] Temporal Database. We're talking about Incarna robotics platform subsequent growth of the Rosetta, it's automation within game model, the platform against Cerebra, which is kind of -- which picks up our internal process, bringing Grid Dynamics as a client 0 for implementations.

What is it all about? Those are not just buzzwords. It's just a way to understand for our clients that there may be a little bit more scarce in the market of clearness what to do. But when you work with Grid Dynamics, we represent basically three key functions.

First, we are domain consultants. So we understand what the customer problems are, and we are tailoring the solutions with that as a important contribution for Grid Dynamics as a mix between Grid Dynamics trained engineers, the standard tools and platform from the market and customized tools, which will bring based on our platform and develop.

The combination of three leads to a few things. First of all, it's a shorter time to implementation for our clients. And second, it moves away from our traditional talent material offering we're putting together contribution based on the planned outcomes, which ultimately leads not only for them to gain momentum and have a better financial return but a high value add for the margin expansion for Grid Dynamics. Those are three ones.

Thank you, Maggie. The next question comes from Bryan Bergin of TD Cowen.

The first one I'll just get a high level. So just with everything that's going on in the market, services, software-based pressure, the whole kind of sets pokey fears that are out there. I want to kind of sanity check it with you first.

Based on what you're seeing in your client conversations and what they're doing in contracting, what -- like what's your perspective as it relates to enterprises increasing their custom build preference versus buy the platform solutions.

And if your clients are demonstrating a rising preference for custom builds, what are like the implications for your dynamics?

So I will start, and then I'll have Vasily to give you a few examples because there's nothing better than to show what exactly happened. So from the high-level perspective, obviously, we recognize that there is a very strong expectation that the cost of implementation will go down.

Then people start throwing some comments. There is a decline of SaaS software companies or [ Ofex ]. There is a decline of IT services needs because everything is going to measure. Well, all these statements are not false. I mean there are more and more tools available in the market. But what's the custom part is, is that creation of the tools and solutions, having our internal platforms makes Grid Dynamics much more efficient to really customize solutions for the individual clients and tests.

And the reason we're doing this because it's very nice from the high-level perspective to look at these all wonderful models, but it's experimentation going to production is quite crazy. And many of the clients are hesitant to throw a lot of money without a clear outcome. And that's where the great comes in with the combination of people, processes and tools. And that's how we believe that even though there is an overall look that over each and looked at there are potential some decline of the needs, the company will agree dynamic needs is actually growing, and I'll have Vasily bring some examples.

Sure. Thank you for the question, Bryan. You are right on point, we definitely see increased demand of our custom-built software. And if in the past, the customers were looking into improvements or enhancing their core platforms, core applications right now, given the overall kind of cost of development is getting reduced by utilizing native environment and DLC.

Companies like Grid Dynamics definitely benefit from this trend by getting involved into implementations and rebuilding of the typically SaaS, I would say, applications as a custom built and more custom tailor solutions for end customers, things like HR systems or travel dashboards and et cetera, which were traditionally work outside of the investment areas for the companies for the clients.

Okay. Okay. That's helpful. And then a follow-up. I'll kind of -- I want to dig in on the growth outlook for the year and unpack it a bit. Anil, you made a comment, you're bullish in your outlook. Just to clarify that comment, are you assuming anything meaningfully changes in the underlying demand backdrop to hit any of these targets? And help us just kind of bridge the 1Q performance here. Is there a build at dynamics or anything seasonal in the first quarter as you think about that first quarter implied growth rate relative to what you're talking about for the year?

Yes, Bryan. Q1 is a very simple story here. It's so -- and also in our time and materials business, T&M, there was fewer working days relative to Q4. So that's -- it's very simple.

Now you're absolutely right. We are positive on how we're looking at the full year. There are two components of it. One is that some of the recent trends in our pipeline growth. Second thing is all the gentlemen that have spoken about on our AI trends, right? I'll let them build up on that. But where we are today, how we look at the year we feel more positive.

And the final thing is that if you look at the range I provided, it's a little wider, right, relative to last year. we made it a little wider because we understand that during the course of the year, there's some positives, there's not so positive. So we kept a healthy range.

So let me be more specific, right? So I think Anil answer a very simple question about Q1, and it's a very substantial reduction of the working days is for. So it's not something like normally happen traditionally here.

But there is a bullish outlook for very simple reasons. The pace of adoption of AI solutions and AI applications by Grid Dynamics customers, clearly outpaces the decline maybe a little bit more hedged retail business. It happens simultaneously and this is no secret because if you look at the rate of growth of our client verticals, you can see to notable changes. It's a tack and more important, the financial vertical, which goes specifically into the fintech and capital markets, which is quite new and growing for us.

So when we look at the total equation, the rate of growth and AI-related businesses. The contribution from our partnerships. The improved our performance in terms of the new type of agreements, fixed bids, performance base, other elements. And on the back side, some of the depreciation of more of traditional paged business, we've been there for years, we came up with bullish but conservative [indiscernible].

And what's the conservative part of that? I think it's very important to understand. We've learned a little bit our lesson from 2025, right? I mean we actually believe we're going to be better than midpoint. But what it means for us? It means for us that in addition to all the facts, we need to understand the revenue dollars which are coming with the customers. And as the business grows, as you know very well, we also deploy our engineering talent across the globe, follow-the-sun strategy. And different regions have different price points and different elements of the business. So as we continue to scale our business, we want to make sure that early on, especially when we're introducing this a little bit variability of Q1, we do not get you guys question, are we saying or not. We are very safe.

Thank you, Bryan. The next question comes from Puneet Jain of JPMorgan.

So given like the recent news flow around Entropic Claude, are you seeing like any changes in your client behavior, increased urgency among your clients to embrace AI? And second, I know like you talked about the gain framework. I know it's built on proprietary as well as third-party tools. So to -- like all these developments like the evolution of ecosystem. Does that raise the bar on what gain can do for your clients in terms of productivity savings?

Very good. Let's start with, again, Vasily the last time to give a little bit more of the multilayer approach. And then from the technology perspective, I think in an comment as well.

Yes. Maybe let me start with again framework. So as you know, we announced it in the middle of 2025. And during the 6 months of 2025, we were rapidly developing this framework and running pilot implementations with our customers.

As you heard in the prepared remarks, we implemented a series of software assets, which became now the part of this platform, which we are offering to our customers.

So I would say in 2026, we see this will be the year of rapid adoption of the game platform across our customers. And in fact, it became the de facto standard approach, which we use for the outcome and output-based engagements. Essentially decoupling billable head count from the revenue growth. So we definitely see performance improvements. We transfer some of that to our customers, and some of that contributes to our improved profitability. Eugene?

Yes. And when it comes to the actual improvements which we are seeing from Agentic core [indiscernible] systems and Claude and of course Entropic kind of others -- of course, many of our customers are embracing it, and we are bringing those capabilities with done together with Rosetta, which is a layer on top of if they are not competing with those Agentic Assistance on the foundation there, but we are making them better stronger and embed our own institutional knowledge in the [indiscernible] systems with every engagement.

And of course, impact of that very much depends on the actual nature of the project. So if you are going into greenfield kind of solution, our gains are immense, like 10 fleets compared to traditional ways because you are creating in an unconstrained environment doing whatever you want.

If you are working in a brownfield project with still well-defined goals, technology modernization and migration you still have a very strong improvement because the agents are 2 years. They are doing things much faster for you. And you see maybe 2, 3x improvements in the performance of the teams. However, when you are coming to the engagement and environments where the majority of the complexity is in the communication or orchestration. This has been -- it's much more challenging to realize the improvements from pure coding and creation of artifact. So it all varies very much depending on the portfolio of our solutions.

Just quickly to add to what Eugene Sue mentioned, I think it's very important we mentioned several times in our prepared remarks as well. I think this transition from T&M based approach to outcome-based and output based. That's -- it's very important to emphasize because this is definitely real. We see that a lot. It happened during the 2025 in transition to 2026. And we see that this year, we will see much more of those -- many more of those engagements going forward. And that's why, as Vasily and Eugene mentioned, our game framework together with verticalized solutions and the platforms that we are leveraging that will be very, very important this year.

Okay. Got it. And let me ask like follow-up to Brown's question on the rest of the year beyond Q1. So based on our math, like it seems like the full year guidance at its midpoint implies like 5%, 5.5% sequential growth beyond Q1. So can you disaggregate that? Like what drives that growth like in terms of like whether it's like you talked about like earlier like the pipeline, billing days and all that. Can you talk about like what drives that 5%, 5.5% sequential growth beyond Q1 to get to the mid-quarter full year numbers?

Anil, make a few comments and, of course, we'll have an to back it up with the numbers. As I mentioned to Bryan, we do very seriously to make sure that we are reasonable but conservative in our [indiscernible] have -- Okay. The pipeline is very robust. And the pipeline which we have right now, not only robust, but it shows a quite opportunity with AI-related products and projects across multiple verticals and multiple plants.

There is always a seasonality, right? So Q2 is better than Q1, and Q3 is better Q2 and then Q4 may have some additional flows like what happens in Q4 last year and all the stuff. But we kind of dissegregate the seasonality and behavior from adoption of AI. And we look at our pipeline as it stands today. So there is a very little assumption, Puneet, that there is going to be some enormous number of white swans or some Hail Mary or something extraordinary grade happens during the course of the year. Obviously, not everything on our books today, but majority is -- and we have a very nice number on tools, accelerated and platforms, which are going to continue to roll out during the year.

So to summarize it, we are not hoping for the numbers. We have a strong pipeline to AI-related projects, particularly in the technology and fintech space. There is a growth in manufacturing, which is cutting quite robustly as well. And we see that adoption, as again, Bryan asked before, of the custom developed solution on a combination of the deployed engineers and train program and our internal tooling brings us much higher acceleration. So the same people, the same trade capacity of the people can have several terms on the execution during the year. That's kind of the high level, but very clear understanding what does that pipeline mean? But maybe Anil will back it up with some number.

Yes. Look, I think the key thing is what Leonard can rate we look at the revenues from a bottoms-up and a top down. And what we have as we go from '25 to '26 transition is this AI factor. And when we looked at that AI revenue kind of bottoms up top down and look at the trajectory, I wish I could give a number, but it's a very healthy number as we go into '26.

That is our foundation for our modeling in '26.

Now when you look at the variations we said, right, we have this wider variation this year. We understand in the course of the year, things can happen. So as you go from the high end to the low end, we bake in some level of conservativeness with some of our clients, especially on the larger side, depending upon how we look at the business today.

But again, this is top-down bottoms-up with some conservativeness, but in '26, the fundamental difference is that we've got this AI trajectory and look at, as Leonard pointed out, look at the fastest-growing segments, TMT and financial verticals. That's the key.

So just again, to put another number, Puneet, because I think it's important. I'll give you a little bit of a prequel, right? So mathematically, it does look a little bit aggressive. But realistically, it's a very unusual quarter to report, right? Is the year-end. So we are in March. So you can suspect that we're probably now numbers in Q1 a little bit better then typically when we present our earnings data a few -- 2, 3 weeks earlier.

So what happened is we see a healthy March. And the impact of this seasonality and less of the working days kind of behind us. So the rate of growth, which you see is based on the lower performance of the first, let's say, 2 months of the quarter. As I was joking, would be lovely to have a Q2 4 months then you can throw all these stuff in the first 2 months of the year, but really, really healthy quarter. So the rate of growth from March on is more, I would say, traditional, which makes us more comfortable with providing the guidance like we are.

Thank you, Puneet. The next questions come from Mayank Tandon of Needham.

Great. Thank you. Anil, you gave guidance on EBITDA for the first quarter, but not for the full year. So I just wanted to check with you, should we expect the same sort of pattern as you mentioned on revenue growth in terms of margin expansion? And do you have any sort of framework on how to think about what the levers are for margins going forward?

Yes. Thanks for that question, Mayank. So as you know, last quarter, we talked about margin expansion in 2026, right? Q4 to Q4, we talked about 300 bps. Within the company, there's several efforts right from internal productivity, right, from geographic optimization, where we're working very diligently on our margin expansion. And that's largely driven by the change of our workforce over the last 3, 4 years, which you all know about.

Along with that, we have investments to Eugene is talking -- Eugene is doing some amazing work in a number of platforms he's rolling out on AI. So it's the balance between the two.

So if you look at our trajectory, Margin expansion, margin continuation is what we are modeling. As the revenue picks up, obviously, you have a little bit more positive leverage there on the EBITDA margins. But the cadence at which these things will play out, you will see in the course of the year, I just don't want to give that level of specificity at this point. But the trajectory should be moving upwards and in line with what we had promised last quarter.

And of course, it's not constant currency situation. So you may want to come.

So the other important thing everyone should understand is that in '25 versus '24 there was a big headwind on FX. So if I look at the cost and revenue on a net basis, that was close to $8 million overall for me, year-to-year. If I look at the last day of '24 and compare what happened on '25. So we're working through that. That's another thing that we're working through.

So to summarize it, I gave you guidance direction of 3% in [indiscernible] Plus. It still stays I hope we can do better than that. There's a lot of opportunities happen. But we're not going to pull the plug and show artificially some numbers related to less investment into Agentic AI or the Physical Robotics AI. These elements are vital for our business, but operational efficiency, the contract efficiency, which we discussed with AI and also distributing workforce more efficiently around the globe. All the three elements. But the driver is fundamentally AI efficiency. That's really the #1 or 3. And I think, you want to...

Yes, I just wanted to comment on the same pretty much along the same lines as I mentioned, right, about fixed-price engagements, right, and outcome-based engagements. That's obviously come typically, with a higher margin. So that's why it's also -- it's part of this program as well. And this year, again, it will be quite substantial.

Got it. And then just very quickly, I wanted to ask about your comments around M&A, Anil. You mentioned that obviously, you have a really good balance sheet and you have the work just to go out and do acquisitions. Are you finding that with the recent market volatility, multiples have come down? Are expectation is a little bit more realistic on some of the potential targets that you might have had in mind?

Yes. Somehow the private companies, they received the memo a little later than the public companies. So the member they finally got, but it took a little time. We are having a good pipeline. Look, we've said that, but I think the number of exclusivities that we have today is as high as it's ever been. It's not done until it's done.

When it comes to valuation, things have come in, they're better than what it was 6 months or 10 months ago. But -- it's still back and forth. Again, remind the most important thing, strategic focus, strategic fit to what we're doing, especially in the AI world that we're entering. That's our bar standards are very high, and we're just not going to buy because we have to buy. We're going to do it if it's strategically fitting.

Yes. I think what Anil didn't tell you it is very obvious we're not buying revenue. This is very, very clear. The relationship we got into the exclusivity with several of the targets, there are very specific in their fashion to address two things. One is the technology components, which we need to add. And the second one is the knowledge of the verticals we would like to be strong with that. So it's not about one size fits all. It's not about just going -- swallowing a big company and report a great number, because usually, it doesn't happen like this. But it's a very specific technology plus verticals.

And it seems as the message you mentioned coming from the from somebody who tells them, okay, now has attained their expectations, I think we're going to be in a better shape because last year, it was the satisfactory.

The next question comes from Logan Chu of Jefferies.

My question revolves around your discussion of kind of moving from labor scaled IP to -- or labor scaled growth to IP scaled growth and kind of the shift from time and materials to outcome based. I'm just wondering what kind of implications that have on your plans for hiring in 2026 and beyond? And then also, where do you think the business model evolves to over the longer term? I mean we have some competitors going all in on kind of subscription-based Agentic delivery, some different competitors saying, no, we don't see it fundamentally changing. I was just wondering where you guys kind of landed on that spectrum?

Okay. So on, I will just say a couple of words, but I think this is a good question for a round table. It's almost like I feel like as a fire chat on the earnings call because there are a lot of elements, which is a very loaded question because you're right, we're kind of the last of the group to kind of present our earnings results and you have there full from everyone telling you something. So it will not be very different. We'll tell you what we think.

So look, the model has changed already. There is no way back and people who will consistently say that, A, nothing changed, or we're going to continue to build the large size of team and more people you have as merrier will probably face some challenges, especially on the large size.

Now I've been saying that for a long time, and it actually works for Grid Dynamics benefit. We're not only a technology-driven company and an innovation-driven company -- we're a nimble company. Our size is fairly optimized. Obviously, there is a place for growth. But we're not having any managed services we're not having some very low-end contracts. And some contracts which were not as progressive or technology contribution, migration all this fashion, they are falling off. And that's why you see this kind of changing of the orders in both ways.

But where we see our model, and I hope [indiscernible] a few examples, is that it's going to be a combination. So it's not the perishable goods of quality engineering. It's a combination of capabilities, trained people and the solutions we have in advance of customer needs, understanding their marketability. We continue to play our role with the partnerships. We understand deeply several key areas, and it can be expert in everything. You try to be expert in everything then you have a very kind of a shallow knowledge and you're going to struggle because you have to fill them all. The bets need to be a bit concentrated even though diversified.

So where I see it's kind of a -- it's a middle ground. One thing which I give you, again, as my input may be a little bit different from others, but it's kind of resonates with our clients very well. The definition of the senior engineer has changed.

So traditionally, the word seniors means the person with many years of experience, they do less here. But today, the definition of the senior engineer means relevancy of the technology competence and a foundational acumen around their own DNA being the moderate age of AI technology. So the age limit changes, but what really changes the depth of the knowledge of people. So the focus of Grid Dynamics is will continue to be supporting the intelligence of information programs. Grid Dynamics University training, green lab training, combination that these fellows also contribute to building on tools, so then they can become much more productive with the clients.

So summarizing my part is that in somewhere in the middle ground, we're bringing the new era of the year engineering the talent, combined with a tooling and a modern world of solving customer problems faster more efficient and combining three elements: people, industry tools and our own platforms. And with that, Vasily, maybe you'll add some...

Yes, just a few comments. Imagine if the customer has a project, let's say, which is provided as a bid for fixed price. And you come and bid for that, let's say, with the pricing 25% to 35% lower than otherwise it would be delivered with a traditional workforce in the T&M manner, let's say, we're like just regular fixed price for the regular engineers.

But actually, you have the productivity of 35% to 45% higher. So that's the clear path for improvement of the profitability, but how do you do that? You implement certain as they'll see new processes on how you develop the software, you deploy a special team, which is very well trained you introduced certain artifacts and assets, which would understand or would fit their vertical we are working in, also understand the coating policies, all the existing coal base set, which would help developers to work to deliver higher productivity. And that's essentially like on the high-level wood game model and what the [indiscernible] is going with.

Essentially verticalized solution, high-performance teams, very well-educated engineers on the modern technology and delivering outcome and output based engagement.

Great. That was very clear. And then I wanted to ask about the partnerships. I know they're 19% of revenues were partner influenced. I just wanted to kind of get a sense of how those partnerships have evolved over time, maybe how you see them evolving in the future?

Okay. So the person who is possible for partnerships, we will bring him in next time, it's a Rahul Bindlish as we're looking okay. we're going to say on poles, right? Thank you for asking this question last because it's actually a very wide part of our growth.

When a few years ago, we started talking about one partnership. We're basically exploring what it means to read the next. And starting with Google, it was great. I mean we have a great experience. We have a great partnership. We have a great positioning of understanding of the modern tools, collaboration.

We have matured significantly ever since. So when we talk about the influence revenue, we're talking about our positioning where we not only contribute to the value of the clients which utilize solutions from our clients and solutions talk about cloud solutions or their modern, large language models or other features. But the elements associated how we are adding our layers, our technology know-how, our technology platforms on the top of their offering, which helps them to penetrate customers faster and helps us to understand earlier what their growth is going to be.

Saying that, we also started to contribute more efficiently to their own developments, on their own products, which is very critical because that's how it drives our business, not only having our partners our vehicle for growth within industry, but is the growing clients themselves.

So from there, we pretty much cover all the hyperscalers. And that's great because it means the customer has a value with Grid Dynamics to get a bespoke solution for the best fit for everyone. And this is good because ultimately, not every offering fits at all, and we are very comfortable to be really good friends with the clients and fair partners with our major hyperscalers.

On the top of it, we're adding more meaningful partnerships and perhaps Eugene can make one of the notable ones because I think it usually gives us a little bit more advantage to fill the gaps on the fast-growing AI implementation where the big guys allow a bit more flexibility for some specialized programs to step in. And since it's going to be probably the last time I speak where Eugene will wrap it up for you. I just want to say one thing which is important, I think, for everyone. It's going to be a good year. We believe in Grid Dynamics. We are having a strong and growing team and I really count and you guys believe in us as we do it ourselves. So thank you with that, and Eugene, please wrap it up.

Yes. Thank you, and this is a great question. And indeed, we -- as Leonard said, we are helping many of our partners to build the value-add components and penetrate new customers and new industries. One notable example is, for example, our partnership with Team Portal, which is our [indiscernible] management system at its core, very robust, very scalable and very powerful. And we apply this system at scale while building enterprise agent platforms, which opened quite a lot of interesting opportunities for temporal to grow into this sector, and we help them to go into major accounts together. And now we enjoy -- it's a good partnership as well.

Thank you, Logan. Ladies and gentlemen, this concludes the Q&A session for today. I will now pass it over to Leonard for closing comments.

This quarter, we demonstrated that AI first transformation is delivering real measurable value. We continue to scale our talent and embed AI driven efficiencies through platforms. By running our AI first operational models, we are proving the same value proposition we advocate for our clients.

We entered the next phase of our journey with a clear road map, a future approved workforce and a steadfast commitment to deliver long-term value for our shareholders. Thank you, and we look forward to updating you on our continuous progress.

Good afternoon, everyone. Welcome to Grid Dynamics Third Quarter 2025 Earnings Conference Call. I'm Cary Savas, Director of Branding and Communications. [Operator Instructions]

Joining us on the call today are CEO, Leonard Livschitz; CFO, Anil Doradla; SVP, Head of Americas, Vasily Sizov; and SVP, Global Head of Partnerships and Marketing, Rahul Bindlish. Following the prepared remarks, we will open the call to your questions. Please note that today's conference call is being recorded.

Before we begin, I would like to remind everyone that today's discussion will contain forward-looking statements. This includes our business and financial outlook and the answers to some of your questions. Such statements are subject to the risks and uncertainty as described in the company's earnings release and other filings with the SEC.

During this call, we will discuss certain non-GAAP measures of our performance. GAAP to non-GAAP financial reconciliations and supplemental financial information are provided in the earnings press release and the 8-K filed with the SEC. You can find all the information I just described in the Investor Relations section of our website.

I now turn the call over to Leonard, our CEO.

Thank you, Cary. Good afternoon, everyone, and thank you for joining us today. Our third quarter revenue of $104.2 million was another all-time high, fueled by AI demand. AI grew 10% on a sequential basis and contributed to over 25% of our third quarter organic revenue. New business resulted in the highest engineering billing headcount.

We remain committed to disciplined capital allocation. I'm happy to report that the Board has authorized $50 million share repurchase program, which we announced in today's press release. This represents about 15% of our company's cash. The buyback reflects our confidence in the long-term prospects of the business and commitment to investing in ourselves. We believe Grid Dynamics shares are undervalued at current market prices, making the repurchase an attractive use of capital.

In Q3, we have the strongest pipeline of new large enterprise logos since the beginning of the year. Customers are regaining confidence and beginning to accelerate their strategic initiatives. We're encouraged by the quality and duration of the new engagements. New programs are multi-quarter in nature and with budget extending well into 2026. This is a substantial improvement from the first half of 2025.

Our partnership influence revenue continued to grow and exceeded 18% of our third quarter revenue. Investments into partnerships are driving faster growth, stronger opportunity pipeline and deeper engagement with new and existing clients.

Grid Dynamics helps customers to build advanced AI and digital capabilities. In addition to the hyperscalers, we are also enhancing our efforts around AI-centric independent vendors or ISVs. In the third quarter, we added 5x more billable engineers than we added in the second quarter.

In the fourth quarter, we expect net billable engineers added to be at the similar levels as in the third quarter. This is indeed a remarkable achievement given year-end seasonal trends. We also grew our average revenue per person by 4% on a sequential basis in third quarter. We continue to rationalize our overall headcount as we align our skill sets and geographies. And this will result in greater efficiencies and higher utilization.

As a result of the strong momentum in the second half of the year, we expect to end the year with a materially higher billable run rate, positioning us well for the growth in 2026. Our 2026 revenue growth will build on the top of this higher baseline, providing a strong foundation for the continued expansion and operating leverage.

I'm also happy to share that we're in the midst of a company-wide initiative to expand our profitability and margins. Over the next 12 months, we expect to improve our margins by at least 300 basis points. We'll achieve our goals through several initiatives that we are currently operationalizing. This includes efficiency improvements with a focus on higher-margin geographies, leveraging enhanced pricing with our AI offering, rebalancing our portfolio of lower-margin business and embracing technologies with our AI-first initiatives.

In particular, we found that the AI tools and frameworks used by our engineers in software development life cycle, or in other words, SDLC are making them materially more productive. In short, they're able to produce more code of better quality in less time. We now have an opportunity to monetize this boost in productivity. AI is the fastest-growing practice in our company, acting as a powerful flywheel for our business.

I'm delighted to share our progress as Grid Dynamics advances its transformation into AI-first company. Our technology vision is clear and structured across 3 horizons: AI-first delivery, Agentic AI at scale and physical AI. This framework guides how we embedded AI into every facet of our operations and service delivery, ensuring our clients receive the most advanced production-ready solutions.

AI First delivery is centered on transforming our engineering and delivery capabilities. This is about operationalizing AI in our core processes. The adoption of AI in SDLC has exploded and our Grid Dynamics AI native service offering known as GAIN is at the heart of this transformation. We're seeing strong adoption of AI First SDLC methodologies with active pilots at major clients, including a leading home goods retailer, a major financial technology company, a prominent health care revenue management provider, a global food distributor and a multi-brand restaurant company.

AI First SDLC fundamentally changes project economics and delivery time lines. It enables us to take on labor-intensive legacy modernization projects that were previously inaccessible, substituting extensive parallel human efforts with specialized teams equipped with AI agents.

The impact on our presales process is equally transformable. Our ability to create full-fledged proof of concepts in hours instead of weeks provides a game-changing advantage, improving conversion rates and accelerate sales cycles. As we deploy these solutions, we see leaders emerging across industries verticals who are driving measurable ROI through new AI capabilities.

While some enterprises are finding success, the vast majority are waiting for commercial off-the-shelf software solutions to become available. It's evident that to achieve meaningful ROI, custom solutions must be engineered for specific business processes, leveraging the foundational capabilities provided by AI leaders such as NVIDIA, Google, Anthropic and OpenAI. This has been the core strength and DNA of Grid Dynamics.

We're a trusted engineering partner that builds from AI-first principles, and this positions us perfectly to help the clients to succeed in this new era. Our second horizon focuses on deployment of Agentic AI platforms for customers and our employees. We are partnering with large enterprises to build bespoke Agentic platforms. This platform-first approach creates significant expansion opportunities.

Our clients engage Grid Dynamics to architect their foundational platform and leverage the expertise to develop sophisticated AI agents for customers and employees, including automated operations with human in the loop. These initiatives drive Agentic customer engagement and enhance decision-making across enterprises.

Our third horizon is Physical AI. It involves integrating AI with the physical world through technologies like digital twins, collaborative robotics and edge computing. The rise of Physical AI is fundamentally transforming the industrial robotics landscape, leading to the replacement of the legacy robotics platform with modern AI-enabled solution. We're advancing our Physical AI initiative through new partnership with selected robotics platform providers.

Our recently announced SmartRay software for robotics world inspection marks an important step forward in our strategy to combine AI with robots. We plan to expand these capabilities further in the coming quarters. The key to success in enterprise AI programs is not just deploying new technologies. It's about having a deep understanding of the business and leveraging technology to solve real-world high-impact problems.

Many companies are realizing that the value of AI comes from rethinking processes, data flows and decision logic around business outcomes rather than pure technology capabilities. This is precisely where Grid Dynamics excels. Our teams combine strong technical expertise with domain influence, enabling us to translate complex business challenges into scalable AI-driven solutions that deliver measurable financial results.

AI projects and engagements serve as a critical entry point for the clients opening the door for larger, high-value platforms and modernization programs. We are seeing a familiar pattern where initial AI engagement such as search or personalization, expands into a broader work across data platform and cloud modernization.

We're capturing a higher share of these initiatives, high-margin projects as clients deploy ROI-driven AI initiatives. We're also capitalizing as the market shifts from experimental proof of concepts to enterprise scale implementations that deliver measurable ROI.

Our game framework continues to gain strong traction with the clients. The model goes well beyond simply layering tools like OpenAI's Codex or Anthropic's Claude Code for the existing engineering teams. The framework rethinks team composition, engineering workflows and best practices to maximize the productivity impact of AI.

The goal is to bring substantially higher efficiency gains than just utilizing stand-alone tools. Over the past quarter, we scaled our expert team dedicated to advancing gain, further strengthening our competitive edge in the AI native engineering space.

And now, I will turn the call over to Vasily Sizov, our Senior Vice President of Americas, to discuss some notable projects highlights from this quarter.

Thank you, Leonard. Good afternoon, everyone. As Leonard highlighted, we are seeing a clear change in customer tone compared to the beginning of the year. Clients who were previously focused on near-term risk management are now taking a more constructive and strategic view, thinking about how to position themselves for growth in 2026 and beyond.

This shift from caution to controlled optimism gives us confidence that the current demand recovery is structural, not temporary. In fact, several of our key customers begin their fiscal year on October 1. Contract renewals we observed and new committed budgets at or above prior year levels indicate maintenance of the momentum.

A significant portion of this activity is centered around AI business cases, initiatives designed to drive tangible operational and financial outcomes. As we mentioned in prior quarters, we are already executing on 2 large-scale platform programs with Fortune 500 clients that are implementing Agentic AI across the enterprise.

We are now seeing a much broader wave of discussions of similar nature and the results to date have been very encouraging. The ROI profile of these AI initiatives looks more attractive than that of traditional digital transformation programs. Unlike gradual multiyear modernization efforts, these AI business cases often target specific pain points with measurable improvements, revenue uplift, cost reduction or conversion rate gains that become visible within quarters, not years. This creates a strong feedback loop as clients see real results, their interest accelerates and demand begins to snowball, which we observe in our pipeline.

With that, I would like to highlight some notable projects from the quarter that illustrate these trends. First, we are developing an AI-driven bug triage solution for a leading multinational technology company. Our approach integrates advanced noise reduction, deduplication and intelligent routing with deep analytical models, custom lock processing and robust domain knowledge base. This combination directly addresses the challenges of engineering operations in bug triage and routing. By automating complex analysis and decision-making, the solution is expected to reduce triage time by up to 70% triage, while significantly improving accuracy and replacing the traditional manual approach for bug triaging.

Second, for a leading technology company, we've developed a system to support compliance with the Digital Markets Act by shifting data processing from server site to on device. Using modern mobileto-edge data processing workflows, the system delivers server source data to user devices for local transformation, a critical requirement under DMA, which prohibits certain server side joints and aggregations. The platform processes billions of records daily across a wide range of business domains and data sets. This initiative has significantly strengthened compliance by enabling privacy preserving non-identifiable consumer data collection at scale.

Third, a leading financial and investment services firm is modernizing its advanced search platform used daily by over 10,000 financial advisers for efficient search of the client data. The legacy interface required navigating nearly 500 filters and understanding of SQL queries and logical expressions, creating complexity and inefficiencies.

The enhanced solution leverages AI and natural language processing to enable advisers to query the firm's databases using simple conversational language. This AI-driven search experience delivers faster insights and an estimated 10% boost of financial advisers' productivity.

And the fourth example, a leading U.S. automotive parts provider had a plan to replace an outdated solar-based search engine with a goal of improving online revenues by at least 3% without disrupting operations. Great Dynamics partnered with them to implement a Google Verdicx AI search solution and successfully accomplished the project with results exceeding the expectations, a 3.33% uplift in revenue per search, generating over $600,000 in just 2 weeks at 50% traffic, outperforming their legacy system by 5%.

Looking ahead, we plan to expand Vertex AI search to B2C and in-store channels with content enrichment and cloud migration.

Now, let me turn the call to SVP Global Head of Partnerships and Marketing, Rahul Bindlish. Rahul?

Thank you, Vasily. Our partner influence revenue has grown to over 18% of total company revenue, underscoring the value of our ecosystem-driven approach. Our partnership framework is purpose-built to advance our mission of enabling enterprises to develop world-class AI and digital solutions.

As AI continues to redefine enterprise transformation, we expect these partnerships to play an even more central role in our growth, driving continued expansion in both our pipeline and market opportunities.

We organize our partners into 2 primary categories. First, platform partners. This group includes the hyperscalers, our core cloud partners as well as leading data and analytics platforms like Snowflake and Databricks. These collaborations keep us at the forefront of modern enterprise infrastructure, ensuring we deliver cutting-edge cloud, data and AI capabilities to our clients.

With the hyperscalers, we are strengthening relationships through targeted investments in AI and Agentic platform capabilities. This includes expanding certifications, earning specialized badges and building new joint solutions, complemented by coordinated go-to-market initiatives such as joint marketing and sales campaigns.

We are expanding these initiatives from the U.S. to other regions, including Europe, LatAm and South Africa. Second, ISV partners. These partners bring deep domain-specific capabilities essential for enterprise transformation. Through these specialized ISVs, we deliver tailored best-in-class solutions aligned with specific business needs.

Within the ISV ecosystem, we have expanded our partnership with leaders in middleware and workflow orchestration that underpins reliable, durable execution for Agentic platforms. Our blueprints for Agentic AI platform, incorporating such middleware developed from real-world enterprise deployments address critical challenges in scaling and managing AI workflows.

We have also expanded our platform partnerships to include NVIDIA. We are actively developing solutions on NVIDIA's advanced software stack, including Omniverse to deliver high fidelity industrial-grade digital twins and simulations.

For example, earlier this year, we launched the Interalogistics optimization starter kit on NVIDIA, enabling retailers, manufacturers and logistics companies to optimize facility layouts and picking paths, boosting warehouse efficiency and reducing labor costs.

With that, let me turn the call to Anil, who will talk about our financials. Thank you.

Thanks, Rahul. Good afternoon, everyone. We recorded the third quarter revenues of $104.2 million, slightly higher than the midpoint of our $103 million to $105 million guidance. On a year-over-year basis, this represents a growth of 19.1%.

On a year-over-year basis, there were roughly 40 bps of FX-related tailwinds. Non-GAAP EBITDA came in at $12.7 million within the higher end of our guidance range of $12 million to $13 million. In the third quarter of 2025, there was a negative impact from FX fluctuations on our costs, both on a quarterly and year-over-year basis.

Grid Dynamics is exposed to a currency basket across Europe, Latin America and India. While we have a natural hedge against some of the currencies and a hedging program with other currencies, the net impact on our EBITDA was approximately $0.6 million or $1.3 million on a quarter-over-quarter and year-over-year basis, respectively.

Looking at performance of our verticals. Retail remained our largest vertical, contributing to $27.8 million of our total revenues in the third quarter of 2025. Revenues in this vertical decreased by 2.1% and 2.9% sequentially and year-over-year basis, respectively. The sequential decline came primarily from a handful of large retail customers, while some of them have returned to growth.

TMT, our second largest vertical accounted for 27.4% of total revenues for the quarter with growth of 13.5% and 18.2% on a quarter-over-quarter basis and year-over-year basis. This growth was primarily driven by our largest technology customers. Finance vertical accounted for 24.6% of total revenues in the quarter. Revenues were slightly up sequentially and grew 81% on a year-over-year basis. The substantial year-over-year growth was primarily driven by increased demand from our fintech customers, along with contributions from our 2024 acquisitions that brought in global banking customers.

Turning to the remaining verticals. CPG and Manufacturing represented 10.5% of quarterly revenues and grew by 3% on a sequential basis and grew by 11.3% on a year-over-year basis, primarily due to contributions from our recent acquisition.

Other vertical contributed 7.4% of total revenues, reflecting sequential decline of 1.6% and 10.5% increase compared to the third quarter of 2024. The year-over-year increase primarily came from customers tied to delivery, service providers and acquisitions.

And finally, health care and pharma made up 2.3% of our revenues for the quarter. We ended the third quarter with a total headcount of 4,971, down from 5,013 employees in the second quarter of 2025 and up from 4,298 in the third quarter of 2024.

During the quarter, we increased our billable headcount meaningfully. That said, we rationalized our overall headcount as we aligned our skill sets and geographic mix. At the end of the third quarter of 2025, our total US headcount was 370 or 7.4% of the company's total headcount versus 8% in the year ago quarter.

Our non-U.S. headcount located in Europe, Americas and India was 4,601 or 92.6%. In the third quarter, revenues from our top 5 and top 10 customers were 40.1% and 58.3%, respectively, compared to 39.8% and 59.2% in the same period a year ago, respectively.

During the third quarter, we had a total of 186 customers, down from 194 in the second quarter of 2025 and 201 in the year ago quarter. The decline in the number of customers was primarily driven by our continued efforts to rationalize our portfolio of nonstrategic customers.

Moving to the income statement. Our GAAP gross profit during the quarter was $34.7 million, or 33.3% compared to $34.5 million or 34.1% in the second quarter of 2025 and $32.7 million or 37.4% in the year ago quarter.

On a non-GAAP basis, our gross profit was $35.2 million or 33.8% compared to $35.1 million or 34.7% in the second quarter of 2025 and $33.3 million or 38% in the year ago quarter. On a year-over-year basis, the decline in gross margin was from a combination of factors that included FX headwinds, higher utilization, lower working time and mix shift from our U.K.-based acquisition.

Non-GAAP EBITDA during the third quarter that excluded interest income expense provision for income taxes, depreciation and amortization, stock-based compensation, restructuring, expenses related to geographic reorganization and transaction and other related costs was $12.7 million or 12.2% of revenues versus $12.7 million or 12.6% of revenues in the second quarter of 2025 and was down from $14.8 million or 16.9% in the year ago quarter. The decrease of $2.1 million on a year-over-year basis was largely due to higher operating expenses and FX headwinds.

Our GAAP net income in the third quarter was $1.2 million or $0.01 per share based on diluted share count of 85.8 million shares compared to the second quarter net income of $5.3 million or $0.06 per share based on a diluted share count of 86.4 million and a net income of $4.3 million or $0.05 per share based on 78.8 million diluted shares in the year ago quarter.

On a non-GAAP basis, in the third quarter, our non-GAAP net income was $8.2 million or $0.09 per share based on 85.8 million diluted shares compared to the second quarter non-GAAP net income of $8.3 million or $0.10 per share based on 86.4 million diluted shares and $10.8 million or $0.14 per share based on 78.8 million diluted shares in the year ago quarter.

On September 30, 2025, our cash and cash equivalents totaled $338.6 million, up from $336.8 million on June 30, 2025. As Leonard mentioned, the Board has authorized a $50 million share buyback, which we announced in today's press release. This represents roughly 15% of our cash. M&A continues to take priority in our capital allocation strategy. We are committed to augmenting our business organically through our acquisitions that strategically enhance our capabilities, geographic presence and industry verticals.

Coming to the fourth quarter guidance, we expect revenues to be in the range of $105 million to $107 million. In the fourth quarter, all our business will be considered organic in nature. We expect our fourth quarter non-GAAP EBITDA to be in the range of $13 million to $14 million. For the fourth quarter of 2025, we expect our basic share count to be in the range of 85 million to 86 million and our diluted share count to be in the range of 86 million to 87 million. Based on our fourth quarter revenue outlook, we expect our full year revenue outlook to be between $410.7 million to $412.7 million. This would represent a 17.1% to 17.7% growth on a year-over-year basis.

That concludes my prepared remarks. We're now ready to take questions.

Thank you, Anil. [Operator Instructions] The first question comes from Puneet Jain of JPMorgan.

So, it was good to see increase in number of billable headcount this quarter, which you also expect to continue into 4Q. Talk to us like about the trends you are seeing for 2026? Like can growth rates next year meaningfully accelerate from like the broad set of clients compared to what you are guiding for 2025?

Thank you, Puneet. It's good to talk at the time when we can be comfortable to discuss the growth. First and foremost, we are the highest billable headcount in the history of the company. The rate of growth has also picked up quite a bit, and we see that going into the Q4. But why we're comfortable looking forward for the next year at this point?

First and foremost, the programs we have recently renewed or we signed for are longer in nature. They're not going on a short duration. They're going on multi-quarters. The second part of that is that the programs are related to the AI initiatives, a lot of technology application, which brings the core of us bread and butter of our business.

The other part, which is important is that we don't get only stuck with the traditional renewals in the beginning of the year because some of our clients now have the sliding schedule for the new fiscal year. So notable clients had their fiscal year starting in October, which means that we are very comfortable to see the growth coming through, again, longer duration.

And finally, as we have told you guys before, we have a number of our top 10 clients who elect Grid Dynamics to be a preferred vendor. It wasn't as evident in the last few quarters because they were a little bit slow on expanding their technology investments. Now they're full swing, and we're taking advantage of that benefiting from being a preferred partner.

Understood. And then a question on Agentic AI, like the benefits to clients from transitioning to Agentic AI-based solutions, it's clear. But perhaps talk to us about the constraints that are limiting adoption? And what will change that? Like could that unlock higher level of discretionary spend among clients for the overall IT services companies next year?

Very good. I would let Vasili talk about some specific cases because obviously, we're in the midst of the big transformation. And there are a lot of talks about what agentic AI can do. And cannot remember, it's our Phase 2 of horizon, the AgenticI is at scale. So, Vasily?

Yes. Thank you, Puneet, for the question. Yes. So Agentic AI and AI in general is the fastest-growing practice for us. So -- and we definitely see the expansion with the business cases, which we already kind of applied during the last few quarters. But the technology doesn't stay on where it is and it continuous evolving. And we are expanding our capabilities to a broader spectrum of business problems to solve. And there are a few notable examples, which we already mentioned during the prepared remarks.

And some of them, for example, apply to the cases where lower skilled employees can be replaced with higher skilled employees in a lower number of, I would say, people augmented by sophisticated AI solutions. And for one of the core clients of us, we are implementing right now the Bug Triage Solution, which is basically assumes a small number of highly expert team augmented by AI, replacing hundreds of engineers of low skilled who are basically don't possess enough skills to requalify and can be really replaced by the more sophisticated processes and solutions.

The next question comes from Bryan Bergin of TD Cowen.

I wanted to follow up actually a little bit on that last question as it relates to the Agentic work and some of the TAM expansion. So particularly this Agentic managed services activity that seems like it's brand new as far as an opportunity for you versus the custom build activity that you're known for. When you think about the work, the Agentic work that you're doing for clients, is there a way to segment how much of it is in this kind of new managed services area versus what would be kind of just SDLC-enhanced Agentic activity? Because I think, obviously, that's a huge market, the IT managed services industry that you could penetrate here in a new way.

Yes. So, thank you so much, Bryan, for the question. I would say, currently, majority of the revenue, which we see are actually related to solving the business cases. As SDLC, I would say, expansion of existing programs and helps to open new accounts, but this is broad in nature. So, it's actually -- it goes through majority of our engagement. So, it's very difficult to discern what exactly would be the incremental gain, I would say. It just fuels overall growth, which we saw in Q3 and was significant.

Yes. So just to comment more specifically, the complexity of scaling the business with the Agentic AI lays in the fact that we cannot just take off-the-shelf program and apply it to the client. When we talk about AI first deployments or more commonly known as forward deployed engineers, it's more or less straightforward.

The Agentic AI unveils a very strong combination of the traditional hyperscalers, with their tools, solutions, their ISVs or some specialty tools and there are -- tools created in-house by Grid Dynamics. And mind you, quite a few initiatives are actually driven by Grid Dynamics to be the client zero, which is now very popular within the company.

So, we train the programs within the company on a business process as an example, and then we carry out to the clients. So, we're expanding rapidly the market because it's a combination of our traditional kind of open sourcing world, but with embracing the partnership and a big players. That's, by the way, one of the reasons I brought Rahul to the call because the success of enrollment into broader base and Gen AI application is driven by how many cable solutions are developed by our partners in conjunction with us doing something in the middle of their preparation for releases.

Okay. That's helpful. That's clear. My follow-up, I'll touch on the numbers here. So just help us with reconciling the 4Q growth view that comes in here a little bit below versus what the prior implied would have been on the fiscal '25 outlook. With all the optimism you're conveying here on billable base on client behavior, is there -- is it just some of the signed work is not immediately starting and it's kind of '26 and thereafter? Is it any client-specific issues? Just anything just on the near-term numbers set up.

Yes. So Bryan, very simple. It's a timing thing. So, there are three layers of this timing. The pickup in ramp, right? We thought it would be a little earlier. It just was a little delayed, but we're getting to the same point. Second thing is that in the year, we had 2 significant clients that had an impact on us. That was about, what, $25 million, $26 million roughly there. And these 2 things -- and the third thing was that if you look at the high end, there was certain M&A also plugged in.

So, when you look at these 3 things, there's nothing structural. As a matter of fact, one of the client, a top 10 client that gave us a little bit of a headache early in the year, that's coming back strong. So, it's all about timing. And that's why when Leonard started off with his opening question, what we are seeing right now going into the fourth quarter sets up very well as we get into '26.

But just to clarify, Bryan, even when we were meeting with you a quarter ago, we could not pinpoint exactly the time of the inflection point. We were reliant first on some time during Q2, then we were not sure. And then finally, the second half of Q3, it happened. So, when Anil refers to the timing difference, if you trace the rate of growth, not from Q1 to Q1, but from Q2 to Q2, you will see significant upside.

And what's very important, again, what Anil said, taking away these 2 clients and some delays we had with them, we're in a fantastic organic rate of growth. Now there always something happens. So, we can't say it will never happen with anyone again. Of course, it happens. But what really carried us out into the more success rate of growth is a broader base of clients. With the application of technology tools we were less dependent on certain variations, especially from our traditional legacy retail and the service business around retail customers as a total.

Okay. Just one clarification. Did you scale that engineering base, the billable base that's up? I know you've got some things flowing through the net headcount from 3Q. Did you scale the engineering side?

So this is very, very simple because when we do reporting, we're trying to follow the same numerical disclosures, right? So, if you look at the number of the billable headcount, it's significantly higher. So, what happened? -- there are 2 things happen. Number one, we're much more aggressive of optimizing the OpEx, right?

The other -- so OpEx is really a big thing. So, we're reducing the bench or anything like that. The second part is it's a small variance of our internship programs. We continue to hire interns, but it happens in the beginning of the quarter. So, when they come to the end of the quarter, it's really not less meaningful, but we really have a robust pipeline of projects and also trained engineers, both from the internship program, Grid Dynamics University, et cetera. But it really -- it looks a bit weird because how you can grow when you have overall headcount. But when you see our engineering headcount, building headcount, utilization, we're extremely positive going forward.

The next questions come from Surinder Thind of Jefferies.

I'd like to start off with a question about the partnership program. Can you talk a little bit about -- when you think about the future of where those numbers could get to, I feel like we've been kind of stuck in the 16%, 17%, 18% range as a contribution or a percentage of revenues. Can you talk about where we are in that process and where you think you can ultimately get to and why the numbers are what they are today?

Thank you for that question, Surinder. A little bit about myself. I was the first salesperson who joined the company more than a decade ago, and we started the partnership program about 4 years ago with the intention to grow our business, increase pipeline, accelerate the sales cycle. And we are doing pretty well on all those parameters.

Now, specifically in terms of percentage of revenues influenced by partnership, we have grown pretty nicely to about 18%. We started off with a goal of getting to about 21-odd percent. But given the growth we have had, I do expect in the long term, we'll end up somewhere between 25% and 30%.

Now, you did make a statement that we have been stuck between the 16%, 18%. In fact, if you look at the trajectory, we are growing from 16% to 18%. If you look at we have also done acquisitions. And a lot of our acquisitions, when they come in, they don't come with partner influence revenues. So, our overall percentage is still growing on a total basis, including acquisitions. Effectively on a dollar basis, our rates are significantly higher.

And then maybe a question on the decision to go with the share repurchase program. Any color there in terms of -- I realize it's not a very large percentage of the cash. But for a growth company, just can you talk about that and what you're trying to signal there? Obviously, I think people recognize valuations are generally depressed, but what's the benefit here?

So Surinder, I think the first and foremost thing is a signal that we're sending to the markets. We believe that we -- our deployment of this capital at these levels is a clear good return on investment. You're absolutely right. We are a growth-oriented company. So, there is a second aspect of our whole story, which is M&A. And we're fully committed towards M&A. This year, we thought we'd close 1 or 2 deals. It took a little longer. But the second aspect of our capital allocation is definitely M&A.

Just to add on this point, I think it's very important. We believe we really passed from the trough. We're in a clear growth inflection point. And we listen to our investors. We understand the market trends, and we believe it's a value which we will coordinately bring to the market, to our shareholders while maintain a very good position on the cash. And also, there's another added factor. We believe we'll generate more cash as the business grows.

And those questions are related to how we're going to improve our EBITDA margin. Why is it important? Why we specifically said something in our commentaries about how actively we're going to do that because we're not just giving away cash. We're making a business-wise decision while demonstrating ability that we'll replenish the cash as we grow with the M&A process forward.

The final quick question here. Just on the margins and the idea of generating 300 basis points of expansion over maybe the next 12 months. I understand this was a year of investment, but can you put that into context why now? Why not continue to invest given all of the change? And is that kind of a onetime step function change that we should then build off of? Or how do we think about the decision to kind of focus on margins at this point in the cycle?

Yes. No, no, great question, Surinder, and that's a perfect question. Look, there are a couple of things that are going on here. The first thing is the timing of it. So, on that one, we believe that the macro is going to be what it is. And we're assuming that we're taking a little bit of a conservative outlook on the macro front and saying that given what it is right now, let's look at the way the business is.

We've also reorganized our headcount. There was many non-repeatable one-off things, whether it is a sudden expansion of geographies, whether one-off discounts as we went through vendor consolidation with some of our big clients and nonrepeatable events, which we've landed with. So, we're taking a look at that.

And the final thing is that, we're embracing these new technologies. So, as we embrace many of these new technologies, we believe that we could get some of the returns. So, where we are on the macro front, on almost like now we are in this new world where we're at 19 countries. Remember, about 3 years ago, we're at 7. We rapidly expanded. So, we're now taking a look at, okay, how should the company look like? What is the optimal model. And we're looking at it on an account-by-account level or region-by-region level.

Let me just add more strategic comment on that. So, there are 3 ways you can manage cost. Number one is on the pricing side. Number two, on the cost. And in this case, Anil alluded to specific regions, which have been affected by the unfavorable exchange rate, and also the transition we had to India as an example. And the third one is a technology investment. And I think you alluded to make sure that we understand how to balance it.

So, answering that question. We are definitely bringing the value to the clients, which is reflected in our new contracts. We're bringing our game model, which helps us to identify the business solutions, which is giving us a bit more run rate on the favorable pricing. The cost is what priority is for Anil to work on. We're not slowing down on a technology investment. Had we slowed down technology investments, it will be a significantly higher number. But my goal in life to bring in Grid Dynamics in the future of the growth with the same or better technology improvements as we had before. So out of those 3 elements, we pursue first 2.

The next question comes from Mayank Tandon of Needham.

Great. I had a couple of quick ones. First is, are you getting any indication from your clients around a potential budget flush in 4Q? If there is upside to your numbers, would that be the main driver? Or are there other factors that could also be potential upside catalysts based on your guidance?

So, Mayank, this is a question that I challenge our teams internally, right, in the timing of it. So, some very interesting things are happening here. Number one is that -- and the gentlemen have alluded to, but let me rehash it. Number one, if you look at all our new deals that we're signing, we're signing at levels, the pricing at that level or higher. Second thing is that all these fiscal lending deals, we're now talking about 2026. So, people who are starting the new fiscal year, that's a very fundamental thing. And the third thing, which Leonard pointed out, see, the year of 2024 going into '25 was all about vendor consolidation. And in some many cases, we went from several to a handful and we've succeeded. We had to give one-off discount. But now they are looking at ramping us, and we're talking about 2026.

So, if you look at my top 10, top 15, which is what, 50% to 80% of my revenues, we're now talking about 2026. That is a fundamental thing. But you're absolutely right, this is something I challenge the team. At this stage, we do not believe it's a budget flush. I'm sure there might be some marginal things, but our fundamental tone here is driven by what we're seeing in '26.

And what's more important from the business standpoint, this is a financial -- very good financial from the business support, we open new programs. When you have traditional what you define as a budget flush, it's unused funds, which are used for some existing projects. This is not the case. We're opening big multi-quarter programs now, which tells you that it would be very difficult for people to appropriate sums just related to the end of the year. So, we're very bullish that this is not a just budget for short term.

Got it. That's very helpful. And then just a quick follow-up on margins. I wanted to just clarify. So, the 300 basis point expansion that you're calling for in 2026, is that gross margins? Or is that EBITDA margins? And then, Leonard, you did go through the levers. Could you just go through them again? I think you went through them a little bit quickly for me, at least. So, I would love to get a little bit more granularity on what the drivers are.

All right. So, I'll let Anil first talk about his favorite topic. Gross margins with EBITDA margin. He loves to talk about it. I'm the one who needs to make it happen. Let him talk about.

All right. Well, look, at the end of the day, Mayank, EBITDA margin -- the gross margin is part of the EBITDA margin, right? So, look, we are looking at the whole P&L holistically, including the cash generation. So, we're looking at the costs. We're looking at the OpEx. We're looking at capitalization. We're looking at every aspect of it. We have at least 300 bps that we're talking about, so by the fourth quarter of next year. We have some aggressive internal targets, and we're balancing that with some of our investments in technologies.

So, I don't want to say it's coming from this, this and this. It's coming from everywhere. And the bottom line is that you'll see expansion. The bottom line, I'm hoping that expansion both on gross margin and EBITDA margin, but we'll see the final numbers.

So now I'm going to repeat what I tried to say to Surinder with a little bit more granularity. So be a little bit patient with me. So, there are 3 elements. One of the pricing increase. The second is a cost optimization. And the third one is the technology investment, okay?

So on the pricing side, there are a couple of elements which are critical. One of the main one is application of our game model. game model for us as we talked about it before. It's all about Grid Dynamics AI application solution enhancements. We talked with Brian about ageentic AI and other elements. So ,some of the programs we're signing now become more favorable. It's not because we're just hammering on the dollars per employee. I mean, that happens or some renewals, but it's not the most effective way.

You actually need to make sure that ROI plays a huge role. And ROI was a bit up term for a long time because when you have a traditional T&M business, what is ROI. Now when you start applying the business solutions and business practices and eliminating a lot of waste on the client side, it becomes tangible. So that's on increasing the profitability of the pricing side.

On the cost side, you asked a very important question. And on the 2 parts, on the gross margin part, Obviously, with -- especially with the dollar versus euro swing, some of the European locations, particularly in European Union zone, become less favorable. So, we're looking at that, how we're going to improve our gross marginality. Of course, you increase the price, but you also look at some of those less favorable locations of the business. That's one of the part.

The second part is, if I mentioned before, if you recall, Grid Dynamics is a client 0. For a lot of internal initiatives, we're testing efficiency on our own business process, and then we transfer to the clients. It would be unreasonable for us to just investigate them and don't take advantage of it. So that's an operational efficiency. that operational efficiency in HR, recruiting, hiring, finance and all these elements of the business, which don't fit traditionally into the COGS, it's all OpEx.

So that's an element. The second big part of element of the cost efficiency. We are putting a lot of effort here because we see that the trend for Grid Dynamics in recent months and quarters after the start of the war, 3 years has been there already, has not been favorable. And it's time for us to tighten the balance from the operational efficiency, but more importantly, the tools. We have a lot more tools. So, it's not just say, okay, you just get rid of this group of people and hire this group. It's a lot of more intelligence.

The third part, we're not touching. -- actually, we're increasing the investment. And again, Surinder asked this question. He wanted to make sure we are not stopping deployment of cash into our technology area. For one way or another, there is -- it's a full P&L. So, whatever we invest in technology is a part of the same EBITDA margin, right? And we invest into all 3 horizons, including the third horizon, which is the physical AI, robotics automation. We work with on the proof of concept and some first project with a very large industrial companies that all takes investments. The partnership work takes investments.

So, we are investing in technology, partnership, 3 horizons of AI, maintaining a strong focus on innovation while looking at the efficiency, both on our COGS and OpEx, and we're pushing the game model to improve the rates on the client side. If I'm still slow, I think we need to talk offline.

The next question comes from Matt Dezort from JPMorgan.

No, no, from William Blair.

Sorry, from William Blair.

William Blair. Matt, you just got bumped up to JPMorgan. Is that an insult or is that a complement?

It's Matt on for Maggie Nolan over at William Blair. I guess to ask another one on margins, Anil, maybe a slightly different way. I guess it doesn't sound like it, but is it -- is that 300 basis points of expansion dependent on your growth reaccelerating next year? Or can you guys expand margins even if budgets and growth remain constrained into next year?

Look, there is some leverage that you get and benefit from a top line growth, right? But as I said in the opening comments, we're not assuming anything much on the macro, not a big positive, not a big help, so to speak. So even if the macro -- even if the demand environment stays the way it was in 2025, we think we -- yes, we are going to expand at least 300 bps.

Right. And I think what's important, again, and you mentioned it before, Matt, we have actually stated this is kind of a bare minimum. So, the market favorable conditions should lift it further. The technology optimization lifted further. It's -- we look with Anil today, it's our 24th earnings call together. It's the first time we're so specific on the margin part because we believe that Grid Dynamics kind of been looked a bit on a negative front from the margins and kind of a little bit tuned down our technology excellence. And we'll look at that and we say, look, we want to make sure that investors truly understand from the granularity how we're going to move forward with a growing business, improving our technology and become really rigorous on our cost-effective initiatives.

Makes sense. Congrats on 24 calls together. Maybe as a follow-up, can I ask about your guys' AI advantaged? I guess, within AI, where does your competitive advantage come from versus your peers? Is it gain? And I guess, is that helping drive the outsized traction and pipeline within any specific verticals where you guys have historical expertise like e-commerce, for instance?

Thank you, Matt. Our history of utilizing AI and in the past, everyone was talking about data analytics, predictive analytics and machine learning. The story started from 2012 when we first started implementing natural language processing for the search engines for the biggest e-commerce and retail companies in the United States.

In 2017, we wrote the first book on AI called Marketing. And we have a long story and a lot of investment into building this expertise, and that's what creates a differentiation for us. So, when the truly AI boom started, we were ready. We had business cases. We had accelerators, blueprints, understanding on how to implement that technology on scale. And essentially, with utilizing LLMs, it's just another tool in our toolbox, which helped us to tackle things which were not possible to tackle before.

So, I would say that, that's the key differentiation. Right now, we are embracing more and more different business cases, different verticals with specific solutions, specific applications of the technology, I would say. And right now, it's difficult to say which industry benefits the most. I would say, everywhere where we are present, we understand how the technology can help and help with figuring out the strategy and the road map for the implementation of AI technology and going more and more into implementing AI platforms, which can help to implement AI technology on the scale of the enterprise.

So just to also look at the bigger picture, Vasily was very good to define some of the application part. You're absolutely right, you have to start from something. And e-commerce and foundational part of the retail business drove us to expansion into the other areas like CPGs, which is very similar in the application side.

But if you look at the recent -- more recent growth, where the applications of AI are becoming more and more prudent, we'll start looking at our growth in a technology TMT segment. We're looking at definitely in fintech. That's been a very successful endeavor for us to expand. And now it's picking up on the industrial side. So, it started from the foundations, but now it expands in the area.

And the second part to your question, look, everybody tells that they have the best position for AI. You can talk to the company which is now valued $5 trillion, or you can call a company which values very little today from what it's supposed to value, which is Grid Dynamics. And everybody tells you almost the same thing, maybe not the jacketed the leather. But the thing is we are very, very laser-focused on the key technology foundational elements, which, as you alluded, was built in the past 12 years, 13 years.

So, we're not going for the super broad-based clients. We're always on a top 1,000 clients in the world. And you can see that, we are tailoring to the programs with the scale in our past experience can benefit the most to the client. So, you have somewhat narrower band at some of the bigger guys who we're competing with. But where we get into the business, we are getting an excellent job and partnering. And that's where the partnership program also expand us, because we started with Google, Microsoft, AWS and NVIDIA. We have those fantastic partners. And when they see the value of Grid Dynamics, then it really speaks for what we are. But the time will tell who is going to be better. So, we're very bullish, but thank you for checking on this point because you need to be us and everybody else honest how good we and others are at AI.

Thank you to our analysts for all your insightful questions. With that said, this concludes the Q&A session for today. I will now pass it over to Leonard, our CEO, for closing comments.

Thank you for joining us today. Our results highlight the strength of Grid Dynamics business expansion and formidable position in AI-driven industry. We have many reasons to be optimistic about our outlook.

A meaningful increase in billable headcount in the second half of 2025 positions us well for the growth in 2026. We focused on double-digit growth in our AI business, scaling our partnership ecosystem, implement margin expansion. All of these underscore our confidence in the long-term potential. Grid Dynamics will continue to deepen its differentiation through technological leadership in the quarters ahead. I look forward to updating you on the next earnings call.