Yirendai Ltd. Sponsored ADR Stock

Good day, and welcome to the Yiren Digital First Quarter 2026 Earnings Conference Call. Before we begin, we'd like to remind you that discussions during this call contain forward-looking statements made under the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995. Such statements are subject to risks, uncertainties and factors that could cause actual results to differ materially from those contained in any such statements. Further information regarding such risks, uncertainties or factors is included in the company's filings with the U.S. Securities and Exchange Commission. We do not undertake any obligation to update any forward-looking statements as required under relevant law.

During the call, we will be referring to certain non-GAAP financial measures and supplemental measures to review and assess the company's operating performance. These non-GAAP financial measures are not intended to be considered in isolation or as a substitute for the financial information prepared and presented in accordance with the U.S. GAAP. For information about those non-GAAP financial measures and the reconciliations to GAAP measures, please refer to the company's earnings press release.

As a reminder, this conference is being recorded. An investor presentation and a webcast replay of this conference call will be available on Yiren Digital's IR website. I will now turn the call over to the company's CEO, Mr. Tang, for opening remarks.

Everyone, and thank you for joining us. The positive trends we discussed last quarter continued to build in the first quarter, marking another important step forward in our transformation. We entered the year with stronger fundamentals in our traditional businesses, while making meaningful progress toward our long-term vision of building an AI-native, multi-industry operating platform anchored by our established fintech businesses.

Operationally, our credit solutions business continued to recover as industry credit conditions improved following a year of challenging regulatory tightening and credit normalization. Through disciplined risk management, AI-powered operational improvements and a continued focus on higher-quality customers, we delivered healthier asset quality, stronger operating efficiency and improved profitability.

At the same time, we accelerated the execution of our All-In AI strategy. Over the past year, we have integrated AI into every major business function, including marketing, customer acquisition, underwriting, risk management, collection and customer service. Today, AI is no longer just a tool for improving productivity; it's becoming a deeper part of how we operate our business.

More importantly, we are expanding these AI capabilities beyond our own operations through internal incubation and strategic investments in AI-native startups, we are building an ecosystem that combines our fintech infrastructure, proprietary AI platform, computing resources and engineering capability with innovative AI applications across high-growth industries.

What makes this strategy different is that our AI capabilities were developed inside real financial services businesses. This gives us practical experience, large-scale data and real business scenarios that can support future expansion into new industries, creating multiple new growth engines while reinforcing the competitive advantages of our existing businesses.

Let me begin with our credit solutions segment. Following a period of regulatory tightening and industry-wide credit normalization, we saw a meaningful improvement in credit quality during the first quarter. This created a healthier operating environment and supported margin expansion. It also builds on the early signs of stabilization we shared last quarter and our AI capability in risk management to give us more confidence that the credit cycle is improving.

Our repeat borrowing ratio reached a record 78% of loan volume, compared with 74% in the same period last year and 77% in the fourth quarter of 2025, reflecting the growing quality and loyalty of our customer base. AI-powered precision marketing continues to improve acquisition efficiency, reducing customer acquisition costs as a percentage of revenue by more than 50% year-over-year.

Credit performance also improved. Our FPD30 plus rate declined to 0.76% in the first quarter of 2026 from 1.16% in the fourth quarter of last year, while our asset recovery rate increased for the first time in 5 quarters. These results demonstrate how our investments in AI are generating tangible business value. They are also improving operating efficiency, strengthening risk management and strengthening also our financial performance.

Looking at our delinquency bucket, the 1 to 30-day rate improved to 2.5%. The 31 to 60-day rate improved to 2.7%. And the 61 to 90-day rate improved to 3.2%, with the early-stage buckets improving meaningfully from their fourth quarter 2025 peaks. Together with our leading credit indicators, these trends confirm that our proactive credit tightening measures are working, and we expect the later-stage buckets to follow as the credit cycle continues to turn.

Turning to our insurance business. Despite continued industry-wide pressure on traditional brokerage commissions, our Internet insurance strategy continued to gain strong momentum. Revenue from Internet insurance business grew by 38% quarter-over-quarter, lifting the overall insurance segment to grow on both a sequential and year-over-year basis for the first time since regulatory reforms were introduced 6 quarters ago.

During the first quarter of 2026, we issued nearly 1 million new insurance policies, representing 135% growth from the same period last year. And the number of insurance clients reached approximately 400,000, up 4.1x year-over-year. These results underscore the stability of our Internet insurance model and its growing contribution to the overall platform.

Looking beyond our core businesses, we believe the emergence of agentic AI represents one of the most significant and far-reaching technology shifts in decades. We are positioning ourselves to capture this opportunity by building an integrated AI ecosystem centered around 3 complementary pillars.

The first pillar is our established fintech platform, including our lending, insurance and other established fintech businesses, while -- which provide recurring cash flow, large-scale application scenarios and valuable proprietary data.

The second pillar is AI infrastructure. We are currently evaluating opportunities to further strengthen our AI computing capability, including the potential consolidation of our existing computing resources to support our growing internal AI initiatives. We are also assessing how these capabilities could over time create opportunities to serve enterprise customers. As this initiative remains at an early stage of evaluation, we are carefully assessing the technical, commercial and capital allocation considerations before making any investment decision. We will provide updates as our assessment progresses and when there are material developments to share.

The third pillar is AI applications. We are incubating specialized AI agents across financial services, including intelligent credit management and insurance assessment, as well as new applications in areas such as education, personal development and entertainment, all of which represent large and rapidly expanding markets with significant long-term growth potential.

Going forward, we will continue to expand these 3 pillars through internal innovation, strategic investments and ecosystem partnerships. Our objective is to build a diversified portfolio of AI-native businesses supported by our proprietary AI infrastructure. At the same time, our established fintech platform will serve as a core enabler of this ecosystem, embedding lending, insurance and other fintech capabilities into AI applications, while providing real-world deployment scenarios, customer access and commercialization opportunities across the platform -- the portfolio.

In parallel, we will continue to invest in the next-generation financial technologies that underpin this ecosystem, including our proprietary AI infrastructure, multi-agent platforms and engineering capabilities, positioning the company to capitalize on the long-term opportunities created by the rapid advancement of AI and adjacent industries.

Now let me walk you through the key AI innovations we have made in recent months. Building on the success of our proprietary large language model, Zhiyu, and the first generation of our multi-agent platform, Magicube 1.0, we recently launched Magicube 2.0. The release marks an important step forward, moving from AI-assisted productivity towards more autonomous enterprise execution.

First, we significantly strengthened AI governance, security and enterprise control. Our intelligent orchestration agent, ZhiNao, serves as the centralized control hub for managing specialized AI agents across the organization by providing unified permission management, governance, auditability and security controls, Magicube 2.0 addresses one of the biggest barriers to enterprise AI adoption and enables organizations to deploy AI agents with greater confidence.

Second, we have moved beyond AI assistance to autonomous AI execution. With the governance framework now in place, our agents are able to execute complex workflows reliably with minimal human intervention. For example, our XuanJi agent can autonomously complete large volumes of operational workflows, reducing cost to serve considerably while improving execution speed, consistency and service responsiveness.

Third, we substantially enhanced enterprise intelligence. Through ZhiNao, Magicube 2.0 seamlessly connects previously siloed enterprise systems, integrates structured and unstructured knowledge across departments and generates more comprehensive context-aware insights. This enables AI agents to produce more accurate, consistent and reliable outcomes across a wide range of business scenarios.

Magicube 2.0 is much more than a product upgrade. It is an enterprise-grade AI operating platform that enables organizations to deploy secure, autonomous and scalable AI agents, driving higher productivity, better decision-making and lower operating costs across both financial and nonfinancial industries. It is becoming the core AI platform that supports both our internal business operations and our long-term ecosystem strategy.

Now let me turn to our AI strategy and the ecosystem we are building to drive our next phase of growth. Over the past 3 years, we have strategically invested in and incubated more than 9 innovative startups. These companies are led by exceptional entrepreneurs with differentiated technologies, strong product vision, and significant market potential across AI and the next-generation technology sectors. Our role extends well beyond that of a financial investor. In addition to providing growth capital, we actively support these companies through talent recruitment, technology collaboration, product strategy, commercialization and business development.

By leveraging our fintech infrastructure, AI platform, engineering capability and public company resources, we help accelerate their path from innovation to scalable businesses. This collaborative model has created a strong strategic alignment between our company and our portfolio founders. As these businesses continue to mature, we believe they have the potential to create meaningful long-term value for both their customers and our shareholders.

Reflecting this shared vision, we have entered into warrant agreements with 4 companies, including the ones we already invested in. While there is no obligation, these agreements provide us with the option to increase our ownership over time. We have the option to take a controlling interest in the future at a prearranged exercise price subject to the achievement of specified operational and strategic milestones. These are staged investment rights and do not constitute current control or consolidation. This structure allows us to participate in potential upside as these companies grow while maintaining disciplined capital allocation and limiting upfront capital commitment. It also provides a flexible and capital-efficient pathway to selectively bring the most successful businesses into our ecosystem over time.

Today I will introduce 2 of these companies, both of which demonstrate how our incubation strategy is translating AI innovation into commercial opportunities. The first company is an AI-native education technology platform focused on delivering personalized large-scale learning experiences. Comparable to leading global AI education platforms, it leverages generative AI to create adaptive learning content tailored to each user's proficiency, significantly improving learning efficiency, accessibility and engagement across language learning and professional skills development. The platform is deeply integrated with China's leading social media ecosystems, enabling highly efficient user acquisition and rapid product distribution. In May, it achieved approximately RMB 2 million in monthly GMV and is growing at a double-digit rate month-over-month, demonstrating strong product market fit and early commercial traction. The company has also started expanding into international markets, creating additional long-term growth opportunities.

Looking ahead, we see 3 primary growth drivers for this business. First, continued product innovation powered by generative AI will further enhance personalization and user retention. AI is fundamentally reshaping the product development cycle, enabling rapid experimentation, faster feature releases and continuous improvements to the user experience, at a pace that was previously unattainable.

Second, we continue to see significant organic growth in the domestic market as AI adoption in education accelerates and the penetration of AI-native learning solutions remains in its early stages.

Third, the company is well positioned to expand its success internationally through overseas market expansion, leveraging its AI-driven platform to efficiently localize content and scale across new markets. We believe this company has the potential to become one of the leading AI-native learning platforms in Asia and an important pillar of our expanding AI ecosystem.

The second company is an AI-native entertainment company focused on building next-generation digital intellectual property. By combining generative AI with creative production, the company is fundamentally transforming how original content is developed, produced and commercialized, allowing high-quality IP to scale much more efficiently than traditional entertainment models.

Its flagship product is a 2.5D anime-style role-playing game that combines tactical combat world exploration and immersive storytelling within a postapocalyptic universe. The game is designed around a highly engaging, character-driven experience complemented by base building and social interaction mechanics that support long-term player engagement, the game has attracted more than 350,000 followers globally, demonstrating strong early community traction and brand recognition.

What differentiates the company is its AI-native content production pipeline. By leveraging generative AI throughout game development and creative production, the company is able to significantly accelerate content creation, shorten development cycles and continuously expand its universe with new characters, story lines and experiences. This capability positions the company to evolve beyond a single game into a scalable, multi-format entertainment franchise spanning animation, music, merchandise, creator content and offline fan engagement, creating multiple recurring monetization opportunities and deeper long-term user engagement.

We believe AI will fundamentally reshape the entertainment industry over the coming decades. The company represents an early example of how AI can accelerate IP creation, deepen user engagement and unlock new business models, making it an important component of our long-term AI ecosystem strategy.

Before I conclude, let me leave you with one final thought. As we see it, AI is not simply another technology cycle [indiscernible] operate how services are delivered and how value is created. Our strategy is not to build a single AI product or participate in a single market opportunity. What we are building is an integrated AI ecosystem that spans infrastructure, enterprise platforms and AI-native applications across multiple high-growth industries.

What differentiates us is the combination of assets we have assembled. Our established fintech businesses continue to generate stable cash flow and provide large-scale commercial application scenarios. Our proprietary AI technologies, computing infrastructure and engineering capability form the technological foundation. Through strategic incubation and investment, we are adding innovative AI companies that expand our ecosystem into education, financial intelligence, entertainment and other emerging sectors.

Together, these elements reinforce one another and create a powerful value chain that is difficult to replicate. We believe this integrated model will allow us to capture value across every layer of the AI economy, from enabling AI infrastructure to powering enterprise transformation, to omni AI-native applications that directly serve millions of users. As each platform company grows, the value of the entire ecosystem increases.

As we move through the year, we'll continue executing this strategy with discipline. We will invest in technologies that strengthen our competitive advantages, partner with exceptional entrepreneurs and selectively bring the most promising businesses into our corporate family. At the same time, we will continue to grow our existing businesses driven by AI-powered lending and insurance, maintain prudent capital allocation and create sustainable long-term shareholder value.

We are still in the early stage of our AI journey, but we have never been more confident in the opportunities ahead. With a stronger traditional business and expanding AI ecosystem and a clear long-term strategy, we believe we are well positioned to create the next generation of intelligent financial and digital services.

Before I close, I also want to thank our entire team whose dedication and resolve through one of the most demanding periods in our recent history made this progress possible. And thank you to our shareholders for your continued trust and support. We look forward to updating you on our progress in the coming quarters.

Now I will pass the call to William to review our financials.

Thank you, Ning. Hello, everyone, and thank you for joining our call. Before I review our financial performance for the first quarter, I would like to point you to our IR website for our earnings release and quarterly IR pack for your reference and additional details.

As Ning mentioned, the first quarter of 2026 was an important inflection point for the company. While our reported results continue to reflect the impact of the industry's credit normalization and the deliberate resizing of our lending portfolio over the past year, our underlying operating fundamentals have meaningfully improved. What we are beginning to see are the financial benefits of the structural change we have made over the past several quarters. This includes more disciplined credit selection, AI-driven operating efficiencies and the continued diversification of our revenue base.

While fintech remains our core business today, we are also laying the financial foundation of new AI-driven growth initiatives that we believe will enhance the resilience of our business over time.

Today I will focus on 5 areas: revenue, credit costs, provisions -- credit costs and provisions, operating expenses, our balance sheet, capital allocation and our outlook.

On the revenue side, the total net revenue for the first quarter was RMB 915.1 million, representing 41% decrease year-over-year, but only 4% decrease sequentially from RMB 957.6 million in the fourth quarter of 2025. This shows an increased stabilization on the credit risk. This was also supported in part by the deferred revenue recognition features of the risk-taking model which is beginning to provide a more stable revenue stream from the legacy assets built up over the past few quarters under this model.

Revenue from the credit solutions business was RMB 795.7 million, down 4% quarter-over-quarter. The relatively stable revenue performance compared with the loan origination reflects the continued recognition of deferred revenues associated with legacy risk-taking assets, which partially offset lower revenue generated from new loan facilitations.

The positive momentum in our insurance brokerage business that we saw in 2025 continued in the first quarter this year as its revenue reached RMB 87.2 million, increasing 4% sequentially and 22% year-over-year, marking another quarter of solid progress following our strategic repositioning of the business towards digital distribution.

Internet insurance now contributes to 29% of the total insurance revenue, compared with 22% in the previous quarter and a negligible contribution year-over-year -- a year ago. The continued migration of consumers towards online insurance purchasing behavior combined with our AI-powered customer acquisition capabilities position this business to become an increasingly meaningful contributor to our revenue mix over time.

Now let's talk about credit costs and provisions. The most significant drivers of our quarter-over-quarter earnings improvement was the normalization of the credit-related provisions. As Ning mentioned, industry-wide credit conditions improved meaningfully during the quarter. Together with our disciplined underwriting strategy, this resulted in lower-than-expected credit loss across our portfolio.

Let's go through the key financial figures associated with the credit risk. The allowance for credit assets, receivables and others declined to RMB 176.4 million, from RMB 302.8 million in the fourth quarter of 2025, a reduction of approximately RMB 126.4 million. This primarily reflects improving portfolio performance and the absence of significant portfolio re-rating adjustment recognized in the prior quarter.

The provision for contingent liabilities was RMB 632.2 million, compared with RMB 1.11 billion in the fourth quarter of 2025, a substantial reduction of RMB 478 million. While provisions remained higher than the same period last year due to a higher proportion of loans facilitated under our risk-taking model, the quarter-over-quarter improvement reflects healthier credit performance and lower loan origination volumes.

Adjusted EBITDA loss for the first quarter of 2026 narrowed significantly to RMB 337 million, compared to a loss of [ RMB 1 billion ] in the fourth quarter of 2025. This was a substantial improvement. The substantial improvement is primarily attributable to the underlying credit recovery in the business and the operating leverage generated by our ongoing AI-driven cost optimization initiatives. We expect these structural improvements to continue supporting earning quality going forward.

During the quarter, we recorded a fair value loss of RMB 89 million, primarily related to the movement in the value of our digital asset holdings. This reflects normal mark-to-market accounting and does not affect the underlying operating performance of our business. Despite that, the net loss improved to RMB 494.7 million from a loss of RMB 868.2 million last quarter. While we monitor the development of macroeconomic and regulatory environment, the continued normalization of credit quality, together with our improving operating efficiency and more diversified business mix give us increasing confidence in the company's projection towards sustainable profitability.

So now let's move to operating expenses. Sales and marketing expenses were RMB 113.6 million, representing a 45% decrease from the fourth quarter 2025. This reflects our disciplined customer acquisition strategy, lower marketing intensity and continued improvement in AI-powered precision marketing. With repeat borrower accounting for 78% of our total loan volume, nearly 4/5 of our lending business now requires minimal incremental acquisition spending, improving the overall efficiencies of our marketing investment.

Origination, servicing and other operating costs declined to RMB 197.6 million from RMB 250.9 million in the previous quarter. This decrease reflects continued operational cost optimization in the insurance business as we transition to digital distribution channels, and it contributes to a higher portion of revenue.

Research and development expenses were RMB 108.9 million, down 10% sequentially but up 27% year-over-year. We will continue to invest in R&D to support our AI ecosystem initiative and monetization of our technologies. This planned increase reflects our deliberate capital allocation toward enterprise AI capability and engineering talent. While these expenditures are recognized as operating expenses under current accounting standards, we view them as strategic investments that strengthen our long-term competitive positions. We expect these investments to continue improving our cost structure and product development capability over time while creating technology assets that support multiple business lines across the company.

In parallel with our internal technology developments, we are selectively investing in AI-native companies that complement our long-term strategy. These investments expand our access to emerging technologies, entrepreneurial talent and new application scenarios while strengthening the broad AI ecosystems which we are building.

General and administrative expenses were RMB 7.5 million, decreased by 26% compared to the first quarter of 2025. The year-over-year improvement in G&A expenses reflect the continued cost optimization within our insurance brokerage operation as the distribution shifts towards more efficient digital channels, together with increasing automation across customer service, operations and collections enabled by our AI platform.

Let's move to balance sheet and capital allocation. Our balance sheet remains strong. As of March 31, 2026, cash and cash equivalents of RMB 2.45 billion and restricted cash of RMB 383.4 million, which, together with financial investments of RMB 507.5 million, brought total liquidity to approximately RMB 3.3 billion.

This strong liquidity position allows us to continue investing in innovation while maintaining prudent approach to risk management and preserving financial flexibility. Beyond our liquidity positions, we have also been steadily building strategic investment that complements our core operating businesses. Under the current accounting standard, many of these investments are reflected either at historical cost or under the equity method, meaning the carrying values may not fully reflect the operational progress or strategic importance to us.

Our objective is not short-term financial gain, but to build long-term strategic partnerships that can enhance our technology capabilities, broaden our AI ecosystems and create additional opportunities for future growth.

As Ning mentioned, some of our investments also include performance-linked warrant arrangements that provide us with the option to acquire more shares that leads to majority stakes at pre-agreed prices if and when any of these portfolio companies achieve specific operational milestones. This structure aligns our capital deployment with the operational progress of our portfolio companies, while preserving balance sheet flexibility. We will continue to evaluate these investments carefully and provide updates as they reach meaningful commercial and financial milestones.

For the financial outlook, looking ahead, we remain cautiously optimistic about the remainder of 2026. First, on credit. The importance in our asset quality -- the improvement in our asset quality has continued through April and May, consistent with the trend Ning discussed earlier. And we expect this to support lower provisioning requirements in the coming quarters.

Second, on growth. We expect the strong momentum from our Internet insurance business to continue as customer behavior increasingly shift towards more digital, 7/24 customer service and on-demand protection solutions. Beyond Internet insurance, our strategy to diversify from traditional fintech to AI-enabled entertainment and learning technologies also creates a very compelling expansion of our growth opportunities.

Third, on AI. Across the organization, AI is delivering tangible benefits in automation, decision-making and operational efficiencies. We believe these benefits will continue to compound as adoption expands across additional business functions. Together with our disciplined internal AI developments and external and synergistic AI investment, these initiatives advance our strategic vision to an AI-native, multi-industry operating platform.

Overall, the company today is structurally different from where it was a year ago. Our earnings profile is becoming increasingly diversified. Our operating model is more efficient. And our technology capability continues to strengthen. At the same time, we are deliberately allocating capital toward AI technologies, strategic investments and warrant positions that complement our existing operations and support our long-term transformations. While this investment remains at different stages of maturity, together they represent an increasingly important component of our capital allocation strategies.

Our capital allocation priorities remain unchanged. We will continue investing prudently in technologies and businesses that strengthen our competitive advantages, maintain a disciplined approach to risk management and preserve the financial flexibility needed to execute our strategies. We believe this balanced capital allocation framework that combines disciplined investments in our core business, internal AI developments and selective external AI investments position us to participate in multiple layers of the AI value chain while maintaining a prudent financial profile.

Thank you. This concludes our prepared remarks. Operator?

Due to time constraints, we will not be holding a Q&A session for today's call. We appreciate your understanding. If you have any further questions, please connect to the IR team of Yiren Digital or Piacente Financial Communications.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.

Hello, and welcome to the 30th Deutsche Bank Depository Receipts Virtual Investor Conference, dbVIC. My name is Zafar Aziz. I'm the DR Investor Relations advisory team at Deutsche Bank. I'm pleased to announce that our next presentation will be from Yiren Digital.

Before handing over to our presenter, some points to note. Please submit your questions at any time throughout the presentation. Finally, all of today's presentations will be recorded and can be accessed via the Deutsche Bank website, adr.db.com. At this point, I'm very pleased to welcome our speaker from Yiren Digital.

Hello, everyone. Thank you for taking the time joining Yiren Digital's Virtual Roadshow Hosted by Deutsche Bank. My name is William Hui and I am the CFO of the company. And we are making a great progress in our AI development and commercialization. These technologies are already helping us building the next-generation fintech platform. Before we get into our presentation, this is the compliance disclaimer.

So let me briefly walk through our company history. So we began digital lending business in 2006. We are listed on the New York Stock Exchange in 2015. In 2019, we expanded into insurance brokerage business through strategic restructuring and acquisitions. In 2025, we facilitated RMB 67.9 billion of loans -- consumer loans to near 4 million individual borrowers. And we also sold nearly 1 million insurance policy. We already sold the insurance policies to 1 million customers.

Since 2021, we have been investing heavily in AI R&D, we have our own GPU computing resource for model trainings. We built our own LLM and successfully integrated into our platform from borrower acquisitions, risk management and customer service. Through our genetic AI platform, Magicube, the Magicube is a platform that incorporates our in-house LLM, Zhiyu and other third-party models plus 6 functional agents that performs designated task because of these AI technologies, we save about RMB 80 million from direct operating costs last year, mainly from lower acquisition costs and more streamlined customer services.

Our outbound call cost as a result of enabling AI in the IVR system was reduced by 84% last year. The service ticket per staff handled increased by 47%. Our AI agent now can handle capital allocation optimization job autonomously. This job was previously conducted by 2 to 5 full-time staff that spent 10 days a month to find the right capital mix. Now AI can do it within 20 minutes. Here are some of the investment highlights. Since the credit cycle in China turned to a cyclical low and was trough in November 2025.

Our early risk indicators have shown our first payment default 30 days are in the lowest since May 2025 when the risk began to shoot up. Our February numbers are near the long-term average level. It indicates our credit solution business has bottomed out. Since the new regulation and the industry overhaul, we are seeing the competition in the industry has eased a lot. The cost of capital and borrower acquisition costs have come down dramatically since the new rule took effect in October. Our revenue base is more diversified with non-rent longer -- higher percentage of revenue driven by our AI innovation income and the growth of our Internet insurance is our first shots of our mixed generation fintech initiatives.

With the turnaround of the fundamentals in changing -- and the change of our revenue mix, we think the company has bottomed at least from the fundamental point of view. Now let's dive into our business. So for our credit solutions business, our flagship digital lending platform, Yi Xiang Hua, incorporates 20 years of lending experience, servicing over 100 million of registered users. In 2025, we have approximately 4 million active users. And 77% of the loan was issued to repeat borrowers. This is a record high. That shows our brand loyalty and the ability to manage risk with existing borrower. With the use of AI, we are able to customize the marketing content to individual customers. The response time for our AI agent to generate personalized new contents is by half a second, with the help of AI we analyzed and approved millions of loans automatically every year.

With the use of AI and improvement in credit markets, our first payment default was down from recent peak of 2.1% to near 1.5%, about the same level as the May 2025. We have done a few measures in the past quarter to mitigate risk we increased and maintained the repeat borrowing amount to 77% of total loan facilitated. The risks from repeat borrowers are more predictable, therefore, that enable us to manage our risk better.

Here are some of the metrics improvement from our AI. Customers, on average, stay longer with the AI agent out-bond call compared to the traditional IVR sales pitch, contributing to a higher sales conversion ratio. Not only the AI agent outbound call is more effective than the traditional IVR, it is also 84% cheaper. Our generative AI model is 50% faster in generating marketing messages to clients. And the conversation is more relevant and personalized. We use AI agent extensively in our debt collection. In 2024, about 45% of our first notice collections were handled by human -- now the number is down to less than 1/4. Our recovery staff, on average, handles 525 cases per quarter and they are 47% more productive.

It is not just the volumes -- they are -- all -- the AI's also handling most complicated cases, as you see on the middle bottom. In 2024, our AI agent only handles the first day delinquent notice. In 2025, it handles 14% to 20% of our longer delinquent cases that require more personalized interaction with the borrower. The AI can, on the spot, [ build ] a repayment proposal or litigation documents, within 5 minutes, while chatting with the borrower. This shows the level of sophistication and efficiencies of -- our AI agent brings to the operation. This is one of the most powerful AI agent we developed last year. The YiQ and the ZhuQue agent. They help us to identify idle capital across our lending entities and predict the capital flow for the next month. It then automatically reallocate capital from the region platform that has a capital surplus to the one that needs more capital. YiQ agent monitor transactions every month, identify the trends that is changed during the period and recommend a capital allocation plan.

ZhuQue agent does the execution, which facilitates the process, including going through the approval process, tracker the fund transfer and so on. We were running this since mid-2025. So far, the capital allocation accuracy has improved 10 -- by 10%, to 66% resulting in lower capital cost by 24 basis points. The processing speed also increases before AI, we used to full-time staff total of 100% hour to make the capital plan with the AI, it takes -- it only takes 20 minutes. It is more efficient, and it allows us to do the allocation on a weekly basis instead of monthly.

Our agent is able to detect when restricted cash becomes unrestricted cash for redeployment. This makes our capital deployments much more efficient, and we are looking to deploy this technology to some of our institutional clients. This is our AI-based risk management process. It starts with a pre-approval, including the KYC. We understand our borrowers' profile through our channels and our record about the borrowers' credit history. Our model consists of over 200 risk models with the data from over 2,000 sources. The model then classified them into 1 of the 10 risk classes based on the risk classes, the system will make credit decisions and pricing automatically. In a post drawdown stage, the AI continues to monitor the borrowers' behavior, predict potential delinquency and propose solution before the loan is due.

On the recovery side, our smart reminder systems can predict the best time to reach borrowers and improve overall recovery efficiency. With the help of AI and improvement in credit cycle, our 30 days first payment default was down to 1.5% from last year's peak of 2.1%. So let's move on to our second business segment. Our second business is insurance brokerage. We are one of the very few platforms in China that has both off-line and online distribution channels that operates nationwide. We recognize customer journey has changed in the past 5 years.

They get a lot more information from social medias and other digital channels, mainly even preferred digital interaction than face-to-face with the live brokers. As the first projects of our next-generation fintech, we leverage our existing channels and funnel down customer demand and identify products that fits the customers and the carriers -- That fit the customer and the carrier have difficultly reaching those customer segments. Our first product is health care insurance products. Our existing customer base or traffic are young professionals. This customer segment has a unique lifestyle for example, they travel a lot for work or for leisure. Carriers are like these customers because they are young and healthy, but hard to reach.

So we determine a few use cases that target these customers and design our own products underwritten by our carrier partners, use our existing digital channel to interact with customers and the result has been very good. Here are the financial results of our Internet insurance business. On a quarterly basis, our fourth quarter 2025 gross premium was RMB 50 million compared to just RMB 4 million in the first quarter of 2025. So that representing the compound quarterly growth rate of 87%.

The gross premium contribution has also been increasing. In fourth quarter of 2025, it was 5.8%, and we expect the revenue contribution from the online channel will reach 50% of overall insurance revenue by 2027. Another interesting point is the online channel has helped reviving our off-line channel as we use the online channel as a low-cost acquisition channel and we cross-sell to our customer, to offline channel for high-margin life or P&C products.

So here are the strategic priority for this year. We will continue to drive AI innovation and operational efficiency. We will leverage our data and AI tools to support new product development as we begin to sell our AI capability as a service to external customers. So we are transforming from an AI adaptive to AI native service offering, starting from the bottom, we have the technology layer with the Zhiyu our proprietary LLM and Magicube, our multi-agent platform. And the sixth AI agent above that use these technologies and other external technologies. These agents can learn, perform and can be sold as separate agents.

For the service layer, we develop functional services such as customer acquisition, credit review and risk management targeting to corporate and consumer segments. And finally, on the business layer on the top, we developed a business model out of these services. And currently, we are turning our credit solution and insurance business to an agentic business. Our growth driver for this year are recovery of our core business as the credit quality improved substantially. Expanding product scope with AI agents, continue to grow our online to off-line insurance model as it is pivotal to our next-generation fintech strategies and finally, exploit the next-generation fintech opportunities.

So turning to financial now, our 2025 revenue was RMB 5.7 billion, so it dipped by about 1.5%. We intentionally slowed down our loan facilitation in the fourth quarter 2025 to weather the credit down cycle. As the technology-related revenue increased in 2025, the revenue contribution from lending reduced and we expect this trend to continue in 2026.

So here's our operating metrics, sorry, here's the operating metrics of our lending business, starting from the left, as you see our first payment default 30 days in January '26 is already down to May 2025 level. When the risk begin to deteriorate, our February and March numbers show the number is near our long-term average as we have increased the loan facilitation effort as the result.

In the middle, the cost of capital decreased by 110 basis points since October 2025 when the new lending regulation was in force. Customer acquisition costs also decreased by 80 basis points year-over-year. These 2 metrics show after the new regulation in the program. Many small players have left the market, resulting in less competitions and also our AI has helped driving a higher sales conversion and more efficient capital allocation.

So this is the breakdown of our insurance portfolios. We have a solid renew book and a new P&C book. The Internet channel has been driving the growth in the past 3 quarters. The growth from the Internet insurance also results in more spill over to the traditional line as the cross-selling is taking effect. On the cost structure, AI has substantial benefits in lowering our cost as the sales and marketing costs as a percentage of revenue decreased from 32% in 2024 to 21% in 2025.

Origination and services as a percentage of revenue decreased for both credit solutions and insurance business. So we are seeing AI has a major effect on our cost structure as the Internet insurance channel has virtually 0 commission to live agents.

So heading to our last slide. So here's our financial highlights. Our 2025 revenue dipped by 1.5% as we decided to slow pace of loan facilitation in the fourth quarter when the risk was high. On the net income side, we changed our lending model to take on a higher proportion of the book under the risk-taking model. We took that approach as the funding pool was stretched during the period of regulation change in the second half of 2025. Many funding partners and guaranteed companies did not want to take the credit risk at that time. We have many high-quality repeat borrowers. So we took a calculated risk to guarantee those loans so that the capital can flow through.

Unfortunately, our accounting standard has a revenue and provision mismatch for this risk-taking model. So we need to take a provision upfront for this self guaranteeing model regardless of the risk but our revenue is amortized over the loan period. As the risk-taking model loan book growth, we took a bigger hit up front, but it results more predictable revenue stream in the future. So we will continue to monitor the market conditions and they adjust the loan mix between the risk-taking and the old model.

So this is the end of our formal presentation. So you may follow us on LinkedIn and X. And now let's move to the Q&A from the audience.

Okay. Let me go through the -- the first question is, given the growing interest in AI agent across industry. Would you consider partnership white label deals where Magicube become the engine behind the platform workflow?

Yes, we are open to different partnerships. But what we want to do is we are not just exporting the technology to the -- to our clients. And we will also work very closely with our clients in transforming their business using the AI agent. So meaning we will also help them how they -- how to use the AI agent to integrate into their workflow, their existing workflow and their business model.

Okay. The second question is, Yiren has shown that AI can cut costs and boost efficiency. Where do -- do you still see an easy win from AI that haven't been fully implemented in the business?

That's a very good question. So we think the Internet insurance will still have a very good growth potential because currently, we only deploy our AI agent to sell the -- what we call the standard product that's sold online, mostly the 1-year health products. So we think as the technology becomes more advanced, we can start rolling out our agent to sell more long-term product like the life or the multiyear property and casualty products, so which has a much higher margin.

Okay. So the next question is, several valuation model shows a very large upside versus the current share price. In your view, what are the 1 or 2 concrete milestones that could trigger a re-rating from here?

That's a very good question. So I think there are 2 factor that will trigger the re-rating, first of all, our existing business. It went through a regulatory uncertainty last year and also together with the down cycle of the -- the down credit cycle. But this year, we have seen so far since January, the risk metric has improved much better than expected. So from our existing business, that's definitely a turnaround. And for our AI business, and as we start generating more technology revenue from these AI products, so we expect -- we are looking to sign more long-term contracts with our clients, so that will enable us to build a more stable revenue stream to get an more predictable cash flow to support our ongoing R&D.

So I think the AI part will help us to not just help our existing fintech business, but we are also developing a new use cases in other industries, so which we hope we will make some announcement later this year to go through that.

Okay. So there are a few audience asking about our international strategies outside of China. As for those who follow us, we started our Philippines business in 2023, and our Indonesian business just started in September 2025. And in the next 12 to 18 months, we will be -- our Indonesian business will start to scale up because we spent the first 6 months understanding the market and collecting more data.

So now we -- and this year, we are in the position to scale up our loan facilitation scale and to generate higher revenue contributions.

Next question. Your presentation mentioned fraud detection blocking large number of high-risk customer every day. How does this AI edge translate to a better terms or larger quota from our fund partner?

That's a very good question. So our fraud detection program actually reads about tens of thousands of documents every day. And last year alone, it saved us about RMB 180 million of potential fraud loss. So I think given those potentials, the funding partner will see our -- definitely, our loan loss has improved a lot and together with the macro, the industry condition with the less competition, so they are more willing to partner with us at a lower cost of capital. So this helped us to improve our margin.

Okay. So one last question. We're running out of time. So the AI leader globally still trade at a premium valuation, what do you think is missing in the current perception of Yiren as a AI power platform rather than just a lender.

I think what we have been building the -- we have been doing our AI R&D since 2021. So I think this 2025 was the year we conduct -- we have launched the products in 2026 and 2027 is the one that we can start monetizing those technologies. So yes, so -- okay, so I think we are running out of time.

Thank you for attending the road show, and then we will -- for those who'll leave your contact address, we will follow up with you with the answers. -- with your questions. But meanwhile, please also follow us on our X and LinkedIn account. So with that, thanks very much.

Good day, and welcome to the Yiren Digital Fourth Quarter and Full Year 2025 Earnings Conference Call. [Operator Instructions] Please note this event is being recorded. I would now like to turn the conference over to Ms. Keyao He, Director of Investor Relations of Yiren Digital. Please go ahead, ma'am.

Thank you, operator. Good morning, and good evening, everyone. Today's call features a presentation by our Founder, Chairman and CEO, Mr. Ning Tang; and our CFO, Mr. William Hui. There will be a question-and-answer session after the prepared remarks.

Before beginning, we'd like to remind you that discussions during this call contain forward-looking statements made under the safe harbor provision of U.S. Private Securities Litigation Reform Act of 1995. Such statements are subject to risks, uncertainties and factors that can cause actual results to differ materially from those contained in any such statements. Further information regarding such risks, uncertainties or factors is included in our filings with the U.S. Securities and Exchange Commission. We do not undertake any obligation to update any forward-looking statements as required under relevant law.

During the call, we will be referring to certain non-GAAP financial measures and supplemental measures to review and assess our operating performance. These non-GAAP financial measures are not intended to be considered in isolation or as a substitute for the financial information prepared and presented in accordance with U.S. GAAP. For information about those non-GAAP financial measures and the reconciliations to GAAP measures, please refer to our earnings press release.

As a reminder, this conference is being recorded. In addition, an investor presentation and a webcast replay of this conference call will be available on our IR website.

I will now pass it on to our CEO, Mr. Tang, for opening remarks.

Thank you, Keyao. Good day, everyone, and thank you all for joining us. In 2025, we celebrated 10-year anniversary of our listing on the New York Stock Exchange. Together, we have reached many milestones. We made a breakthrough in our AI innovation where we completed regulatory filing of our own large language model, Zhiyu. In the second half of the year, we released our first multi-agent platform, Magicube. With support of these AI tools, we incubated our Internet insurance business, which has achieved strong growth quarter -- strong growth quarter-after-quarter in 2025.

2025 was also a year that demanded the best of us and our team delivered. Heightened credit regulations and industry-wide deterioration in credit quality created significant pressure across our business. Yet we navigated these headwinds with discipline and operational resilience. Equally important, we enter 2026 with growing confidence. Our next-generation fintech platform is gaining meaningful traction and validating the strategic investments we've made.

I'm deeply grateful to our entire team for their dedication and resolve through one of the most challenging periods in our recent history. The rapid advancement of AI is fundamentally reshaping the industries we operate in, and we believe we are uniquely positioned to lead that transformation. Our years of deep vertical expertise in credit facilitation and insurance brokerage, combined with the AI infrastructure and agent technologies we've purposefully built give us a differentiated foundation to reimagine our business ecosystem, accelerating growth and unlocking new avenues of innovation.

Amid these challenges, we made meaningful progress on the 2 strategic priorities that will define Yiren Digital's next chapter, the continued scaling of Internet insurance distribution as our second core growth engine and the accelerating integration of AI capabilities across our business operations. Both are delivering results and both give us confidence in the trajectory ahead.

For years, we have applied our proprietary AI capabilities to continuously analyze our platform data, systematically searching for where our next growth opportunity lies. That process of disciplined discovery led us to a clear and compelling insight. Our users demonstrated strong validated demand for online insurance products, demand that was underserved and ripe for a technology-driven solution.

In the third quarter of 2025, gross written premiums generated through our Internet insurance distribution business surged by 206% quarter-over-quarter. This strong momentum continued in the fourth quarter with another 95% quarter-over-quarter growth and the revenue contribution to the segment had reached 22% in the fourth quarter.

2025 was also a landmark year in the comprehensive build-out of our AI infrastructure, where we closed the gaps and reached significant milestones. Following the regulatory filing of Zhiyu, our proprietary large language model in April, we launched Magicube in October, our internally developed agent integration platform purpose-built for enterprise scale AI deployment. Magicube is the connective infrastructure that enables large-scale coordinated deployment of multi-agents across every critical function of our credit lending business from sales and risk management to capital planning, compliance and customer service. With Magicube in place, we have laid the foundation to automate processes with AI-driven agents throughout our operations. A transformation that we believe will fundamentally redefine how Yiren Digital operates and competes in the marketplace.

The depth of our AI integration is best reflected in its financial impact. In 2025, AI-driven optimizations generated cost savings exceeding RMB 80 million, driven by the deployment of AIGC for marketing and AI-assisted outbound customer service, capabilities that have structurally reduced our dependence on both external vendors and internal headcount and better cost and capital efficiency. The operational impact of our AI deployment is best illustrated through concrete examples. Response times for our real-time AIGC-powered customer service [indiscernible] generation were cut by more than half from 1.2 seconds to under 0.6 seconds, delivering measurably smoother customer interactions at scale.

In a particularly compelling demonstration of our internal AI capabilities, our R&D team rebuilt our IVR system entirely in-house, decreasing our dependence on an external vendor and reducing the cost per call by 84% from RMB 0.95 to RMB 0.15. Meanwhile, our AI-powered intelligent routing 2.0 system brought a step change in productivity to our fund management team, replacing legacy Excel-based workflows with an intelligent natural language interface driven by our 2 proprietary AI agents, YiQ agent and ZhuQue bot, fundamentally modernizing how our team operates day-to-day.

These technological advancements are not just improving how we operate, they are redefining who we are. Our AI-enabled capabilities across intelligent marketing, smart capital management and advanced risk control have strengthened our ability to deliver technology solutions to the broader credit industry. Revenue from technology-driven services, including networking, marketing and technical support has grown significantly year-over-year, validating the commercial potential of our AI capabilities beyond our core business. We are now accelerating this growth to transform the company from a fintech platform into an AI-native company for multiple industries.

Finally, I'd like to review the performance of our credit solutions business against the market backdrop in 2025. In the fourth quarter, we facilitated RMB 12.0 billion in loan originations, moderated by 22% year-over-year and 40% quarter-over-quarter. The moderation reflected our financial discipline when credit environment was difficult. We focused on higher quality credit during the quarter, which led to reduction in loan facilitation activities. For the full year, however, total loan facilitation reached RMB 67.8 billion, up by 26% from RMB 53.6 billion in 2024. As of December 31, 2025, the cumulative number of borrowers we have served exceeded RMB 14.3 million, representing a 16% increase from approximately RMB 12.4 million at the end of 2024.

During 2025, we strengthened our customer analytics and operational management with a particular focus on maximizing the lifetime value of high-quality repeat borrowers. At the same time, we maintained a prudent approach toward new customer acquisition. Through enhanced data analytics and more refined customer segmentation, we prioritized the management and engagement of high-quality existing borrowers. As a result, our repeat borrowing volume remained high at 77% in the fourth quarter of 2025 compared to 65% in the same period of 2024. Meanwhile, the average loan ticket size on our lending platform increased from RMB 8,000 in the fourth quarter to RMB 11,500 in the fourth quarter of 2025. These operational strategies allowed us to effectively control customer acquisition costs while retaining higher-quality borrowers with deeper credit insights and stronger brand trust.

The quality from the legacy assets came under pressure in the fourth quarter with the delinquency rate reaching a cyclical high in October. Our 1- to 30-day delinquency rate for fourth quarter reached 3.4%. The 31- to 60-day rate was 3.0% and the 61- to 90-day rate stood at 2.8%. These levels are in line with industry trends and the macroeconomic environment.

During 2025, assets under the risk-taking model nearly doubled, which contributed to an increase in our guaranteed service revenue as a result of changing credit requirements by our partners. Encouragingly, our lending -- our leading risk indicators are beginning to turn. Our first payment default rate FPD30 for loan delinquency over 30 days has been on a declining trend since October 2025, recently approaching the levels observed in the first half year of 2025. We believe these are early but meaningful signals that the credit cycle is gradually turning, and we expect a broader easing of the credit environment to support continued improvement in both industry conditions and our own asset quality metrics, giving us well-funded confidence in our ability to deliver disciplined and stable operations in 2026.

On the institutional funding side, we secured wide list status with 29 institutional funding partners as of the end of 2025, and this number continues to grow in the new year, reflecting recognition of our risk management capability and financial discipline by our partners as well as less competition in the market and the new regulatory framework. As the industry digests the impact of the new regulations and the market consolidates, we are confident that leading highly compliant players like us will benefit. In overseas markets, we expect to gradually expand our operations in the existing Philippines and Indonesian markets while maintaining prudent financial discipline and a clear focus on profitability. We look forward to showing more results in the coming quarters.

As mentioned earlier, our traditional insurance brokerage business, which is predominantly anchored in a traditional sales network has found new direction of growth. Amid the regulatory headwind on commission rate and the macroeconomic challenges in the fourth quarter, gross written premiums of our insurance brokerage business reached RMB 860.1 million, down 22% year-over-year, while full year premiums reached RMB 3.7 billion, a 17% decline from 2024. However, the composition of the revenue and the premium has changed significantly as contribution from Internet insurance business increased rapidly in the past few quarters, largely filling up the gap from the traditional line.

Our Internet insurance business has delivered a meaningful expansion in both customer base and policy volumes, reinforcing our conviction that Internet insurance represents a sustainable and scalable second growth engine for Yiren Digital. For the insurance brokerage business as a whole, at the end of 2025, we had served over 2 million insurance clients, up 33% from 1.53 million at the end of 2024. New policies issued reached 2.3 million, a 25% increase from 1.8 million in 2024. As Internet insurance continues to contribute more to our total brokerage revenue in 2026 and serves as a low-cost customer acquisition channel for the entire platform, we are confident that our insurance business will successfully turn to both growth and profitability.

To summarize, 2025 was a year that demanded resilience and revealed opportunity. We navigated one of the most challenging credit environments in recent history while simultaneously advancing our transformation into a next-generation fintech driven by AI. The explosive global growth of AI is reshaping customer -- consumer credit, insurance and industries far beyond, and we intend to be at the forefront of that transformation, not merely a participant in it.

We are actively building towards that future, incubating AI native business models, developing technology-driven revenue streams from our credit solutions and reshaping our insurance brokerage business by fully integrating our online and offline capabilities as the cornerstone of long-term growth. Encouragingly, leading indicators increasingly signal that the worst of the credit stress cycle is behind us and our core lending business is embracing recovery with renewed momentum.

As we enter 2026, we are optimistic about the recovery of our core business. We are confident in our strategy and commitment from the team that delivered through one of the most demanding years. Our AI foundation has been laid, and we continue building it.

With that, I will now pass it over to William, who will provide more details on the financials for this quarter and the full year.

Thank you, Ning. Hello, everyone. I will be walking you through our financial performance for the fourth quarter and full year 2025. Please refer to our earnings release and IR deck for further details, both available on our website.

This quarter reflects continued progress across several of our key strategic priorities. First, our investment in AI are beginning to translate into tangible outcomes. We achieved direct net cost savings of approximately RMB 80 million, driven by improvements in areas such as high sales conversion, customer service automation and risk management efficiency. This figure excludes other business benefits from AI such as avoidance of fraud losses, savings from staff training and other indirect cost savings because of AI.

In addition, our proprietary AI technology is beginning to generate revenue in new business within the credit solutions and Internet insurance segment. Second, our Internet insurance business continues to gain momentum. During the quarter, we recorded gross written premiums of RMB 50 million, representing 95% quarter-over-quarter growth. The annualized premium reached RMB 267 million in the fourth quarter, representing 36% growth quarter-over-quarter, up from a negligible amount in the fourth quarter of 2024. The revenue accounted for 22% of the revenue from our entire insurance segment in the fourth quarter of 2025. We expect this revenue contribution to continue to grow and take a bigger revenue share in 2026.

For the credit solutions business, 2025 was a unique year. We began with a very good growth momentum, seeing a 43% growth in loan facilitation volume in the first half of 2025. However, we subsequently faced a downward trend in the credit cycle alongside with regulatory changes. In the second half of 2025, we shift our strategic priority to credit quality over loan growth, resulting in a 22% year-over-year contraction in our loan volume.

Having said that, we are seeing early signs of turnaround in our credit cycle. Key credit metrics have improved. The 30 days first payment delinquency or FPD rate peaked in October 2025 and began to stabilize and trend down in November 2025. Figures for December 2025 and January 2026 have improved more than expected. The delinquency rate in February was 38% below the peak, which is already back to the May 2025 level when credit quality began to deteriorate. However, as a reminder, there is typically a lag of 1 or 2 quarters before these improvements are fully reflected in our financial results.

Overall, we remain focused on maintaining strong balance sheet with cash positions of RMB 3.3 billion, while continuing to invest in AI capabilities and high-growth opportunities. We believe this balanced approach positions us well for sustainable long-term growth.

Turning to the key financial figures for the fourth quarter and full year of 2025. Total revenue for the full year 2025 was RMB 5.72 billion, representing 1.5% decrease from 2024. The decrease was a result of prioritizing credit quality over loan growth in the second half of the year as we tightened our credit policy in response to a challenging credit environment. Full year loan facilitation volume was RMB 67.8 billion, representing 26% growth comparing to the full year of 2024. This growth was driven by strong performance in the first 3 quarters partially offset by a contraction in loan volume during the fourth quarter of 2025. Our guarantee services also saw significant growth with revenue reaching RMB 612 million in the fourth quarter of 2025, up nearly 196% year-over-year as we shift more loan origination to a risk-taking model during the year.

Regarding credit quality, our 31 to 60 days and 61 to 90 days delinquency rates reached 3% and 2.8%, respectively, in the fourth quarter, while the 1 to 30 days delinquency rate reached 3.4% in the fourth quarter of 2025. This reflects the higher risk environment and in response, we have tightened our credit policies. Our upgraded AI-driven risk management system is enabling us to more frequently and effectively assess and mitigate risk across our portfolio.

Looking at the same metrics on a monthly basis, the delinquency rate peaked in October and gradually decreased in December 2025. For instance, the 1 to 30 days FPD rate decreased by 38% from October 2025 to January 2026. For the customer acquisitions, our AI models have enhanced our ability to understand customer behavior and execute more effective precision marketing strategies to drive higher sales conversion. As a result, customer acquisition cost as a percentage of total loan facilitation volume declined by 80 basis points to a record low in the fourth quarter compared to the same period in 2024.

In the insurance brokerage segment, our gross written premium decreased by 22% year-over-year to RMB 860 million in the fourth quarter of 2025, and the full year gross premium was down by 17% year-over-year. The decrease was due to premium from the traditional channel which decreased by RMB 290 million. That decrease was partially offset by RMB 50 million increase from the insurance -- from the Internet insurance. For the fourth quarter of 2025, revenue from the overall insurance brokerage segment was RMB 84 million compared to RMB 106 million in the same period of 2024. The Internet insurance revenue contributions accounts for 22% of the total segment revenue in the fourth quarter and 14% for the full year of 2025. It has become a significant part of the business.

The integration of our online and offline channel, combined with our AI-driven sales and servicing capabilities, enhance the overall customer experience while supporting more competitive customer acquisition cost structure for our traditional business. This integrated approach also creates opportunities for increased synergies across channels.

We also recorded technology-driven marketing service revenue in the fourth quarter as we are transforming our organization into an AI solution platform company. We look forward to presenting you more details in the coming quarters as these services scale.

On the expense side, sales and marketing expenses in the fourth quarter of 2025 decreased by 31% year-over-year to RMB 206 million. This is attributable to lower origination volume, lower acquisition cost for new customer driven by AI and an increase in our repeat borrower ratio to 76% through the year. The overall customer acquisition cost as a percentage of loan volume decreased by 80 basis points in the fourth quarter of 2025 compared to the same period of 2024. It was a record low, reflecting less competition in the market as some players exited the market following the new regulation and also our AI marketing strategy, which was driving a better customer acquisition efficiency.

Research and development expenses decreased by 26% year-over-year to RMB 121 million in the fourth quarter of 2025. This was due to a high base effect from the expense of our credit analysis system development project in the second half of 2024. The full year R&D expenses were RMB 407 million, representing 1.3% decrease from 2024. With the innovative AI tools we are building more for less, we will continue to invest in talent and AI infrastructure to enhance the overall productivity of the R&D team.

Origination, servicing and other operating costs increased by 27% year-over-year to RMB 251 million in the fourth quarter of 2025. This was driven by increased commission rates for asset recovery services to boost collection incentive during a challenging credit environment. Full year origination and servicing costs decreased by 11% to RMB 786 million, driven by decrease in insurance brokerage business costs, along with the increased AI automation as over 81% of our first payment delinquent cases are being handled by our AI agents.

General and administrative expenses for the quarter increased by 4.8% year-over-year to RMB 44 million. We have imposed tighter cost control to lower the expenses further. The allowance for contract assets and receivable for the fourth quarter decreased by 46% year-over-year to RMB 296 million, driven by higher receivables from guaranteed services and financing services amid industry level higher risk profile of assets.

Provisions for contingent liability this quarter increased by 343% year-over-year to RMB 1.1 billion, reflecting the growth in loan origination volume under the risk-taking model, which grew by 48% year-over-year. Under the current accounting standard, we are required to recognize provisions for contingent liability immediately upon loan origination under the risk-taking model, while the corresponding revenue is amortized over the loan period. The increasing proportion of the risk-taking model loan volume has -- will continue to have an accounting impact on our earnings in the coming quarters.

As previously noted, accounting standards give rise to a timing mismatch that results in a near-term earning pressure when we taking model loan volume growth because standby guarantee liability is recorded on the balance sheet at loan inception. This liability will be amortized to become guaranteed service revenue over the guaranteed period in the future, where the provisions for the associated guarantee-related contingent liability and standby guarantee liabilities are recognized upfront in accordance with GAAP, resulting in timing mismatch for revenue and cost, this timing mismatch is expected to normalize when the loan balance under the risk-taking model stabilize when the amortized revenues from the legacy assets balance out the provisions from new loans.

For the fourth quarter of 2025, GAAP net loss amounts to RMB 882 million, largely due to higher accounting provisions driven from the guarantee business, as mentioned. The moderation in performance of the traditional insurance business and RMB 109 million fair value loss on the crypto assets. For the full year of 2025, the GAAP net income was RMB 14.5 million. To match the revenue and contingent liability accrual after adjusting for revenue from the stand ready guarantee liabilities, our non-GAAP net income for the full year 2025 was about RMB 834 million. So regarding our cash flow, we recorded a net cash outflow from our operation of RMB 198 million in the fourth quarter of 2025, but our balance sheet remains strong with cash and cash equivalents of RMB 3.3 billion as of December 31, 2025.

Looking ahead to 2026, our non-lending business will continue to drive our revenue growth, while our credit performance continued to improve on a sequential basis. As this is a conservative forecast as our delinquency figures are improving more than expected, we may revise our forecast during the year. Overall, we are optimistic about the business as the core business has shown signs of recovery. Our Internet business has become a significant growth contributor, and our AI platform engine is starting to deliver results.

So that's the end of our presentation. Operator, back to you.

[Operator Instructions] And our first question will come from [ Connie Gu ] with [indiscernible].

And my question is about AI. You mentioned that internal AI transformation has brought significant cost savings to the company in 2025. So looking to the longer term, do you expect further cost savings or a broader potential for AI application scenarios? And when we compare in-house developed AI agents to the third-party ones, what are the specific advantages? And how do you view the security of the popular AI tools lately like open?

Thank you for your question regarding AI. And actually, let me, explain more, talk more about our AI strategy. And I think it's extremely important because we are, as I reported earlier on, redefining the company. Yes, previously, it's a fintech company utilizing technology to do better finance credit work to begin with. Then we included insurance. But in the future, it's going to be an AI agent, AI native company, not only for credit and insurance subsectors, but also for more financial services subsectors and a few select industries in the coming couple of years.

So basically, it's going to be a different value proposition evolving from our past. Let me explain more. Yes. So when we first started to utilize AI. It was more like a tool for cost savings to do our existing processes better, cheaper. Yes, AI as a tool. That's like in like 2023, 2024, but from last year and even more so this year, you just mentioned the [ OpenClaw ], AI is now a colleague is now a person, yes, a worker. So that means we are going to do businesses differently. We're going to reengineer our business processes for credit and insurance existing businesses. And at the same time, because our technologies, our AI capabilities have been well tested, proven in this heavily regulated demanding super tight security standard industries, sectors, our AI capabilities, our agents can be utilized for other financial services needs, subsectors and going beyond financial services subsectors to more industries. So this is the strategy. This is the development process, yes.

So going forward, we're going to do AI more and more for our credit, for our insurance businesses. But at the same time, we'll look for more subsectors in financial services and new verticals beyond financial services to leverage our AI capabilities, proven capabilities, yes. So this is the strategy we have.

And my vision is after 1 year, 2 years, 3 years, Yiren Digital will be a different company. It's not totally away from our traditional businesses. They will do -- we will do like credit, we'll do insurance. These are great applications for AI. But at the same time, we're going to do more. Yes, there are better also subsectors for AI applications, agents and growth for us going forward. So this is the strategy we have in mind, and we are executing. Thank you.

Yes. Just to add to Ning's comment with the numbers, in 2025, we already achieved a cost saving of RMB 80 million, and that is on top -- and those are the direct -- just the direct costs, and that's on top of other indirect cost savings such as the avoidance of fraud losses, which was approximately RMB 180 million last year and also other costs like the staff training and office space and all that. So I think just to add on to Ning's comments, we are transforming the company from just being using the AI to save cost to using the AI to generate revenue. So the -- what AI will help us is it will reduce our time to market with the technologies and also the analytics that will help us to identify new business opportunities.

The next question will come from [ Wang Yang ] with [indiscernible] Securities.

[Interpreted] Since the new loan facilitation regulation issued in October 2025, the industry has generally experienced a significant impact. Has the company seen any improvement in this effect so far? How do you expect the industry risk environment to evolve over the course of the year?

Okay. Thank you. Thank you for your questions. Based on our credit performance metrics, our risk level peak in the last October and now showing signs of recovery. The new industry regulation had a short-term impact on us, our funding partners and our peers. We believe the industry has already adapted to this short-term impact, position itself for better long-term development. Our January FPD30 and DPD30 metrics, which track 30 days delinquency rate have dropped by 38% to the level seen in May 2025 when this cycle began. When the new regulation took effect in October, our cost of capital -- sorry, since the new regulation took effect in October, our cost of capital has decreased by 93 basis points. Meanwhile, our customer acquisition cost as a percentage of loan volume continued to drop by another 0.8% to a record low now.

So indicating after the new regulation, the competition is -- has been eased, and we view this as a positive signal. And our balance sheet remains solid, providing the financial strength to manage potential risk as these improvements continue to flow through the business. But overall, we remain confident in the long-term fundamentals of our business. We think the new business will make the industry healthier. Thank you.

The next question will come from [ Yulong Yu ].

[Interpreted] I have noticed that the company's Internet insurance distribution business has demonstrated strong breakout growth. Could you elaborate on the development targets and strategic priorities for this segment in the new year? Additionally, compared with traditional insurance distribution models, where do you see our key competitive advantages are?

Okay. Let me take the first crack and yes, William can add to it. The insurance -- internet insurance business market potential is very big. Yes. And you may well remember that our credit facilitation business actually was quite offline several years ago. And then we successfully moved it to online to, yes, digitally transform the business. That was absolutely necessary, the right thing to do, bring us growth opportunities. And the same is happening for our insurance brokerage business, but not exactly the same, let me explain.

Well, more and more businesses are moving online, yes. So the online part will be bigger and bigger contribution to our insurance business, top line, bottom line. And the same is happening as the credit business going from offline to online. The difference is we will still have offline part, but that offline part is also going to be more and more kind of like the so-called offline and online, meaning our offline colleagues will do more and more online activities like live streaming, like WeChat, Douyin kind of applications. We'll do that more and more. So the offline part will be also more and more effective. And as you have seen, the online part, the purely online part is showing great potential, super fast growth. And that's also very promising.

So going forward, the insurance brokerage business will have this, yes, high growing like online part and also a more efficient like offline part kind of being offline, online combined model. Yes. So this is the vision we have for our insurance business. And William, do you have anything to add?

And by the way, I'd like to add something why like our online Internet insurance business is growing, yes, so fast, much faster than our credit business transforming from offline to online because pretty much all the tools have been built for the credit business, the analytics, the AI agents, capabilities, so on, have been built. So it's a much faster acceleration process. And the same logic goes for what I just mentioned, us moving to other like verticals, other industries, the same kind of AI infrastructure, the agent capabilities have been built. Of course, we need to add new kind of like vertical domain expertise. That's also essential. But to begin with, the technology platform capabilities have been built. So it's a much faster, much accelerated process. Thank you.

And that will conclude our question-and-answer session. If you have any further questions, please connect to the IR team of the Yiren Digital or Piacente Financial Communications. The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.

Good day, and welcome to the Yiren Digital Third Quarter 2025 Earnings Conference Call. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Keyao He. Please go ahead.

Thank you, Operator. Good morning, and good evening, everyone. Today's call features a presentation by our founder, Chairman and CEO of CreditEase, our CEO, Mr. Ning Tang; and our CFO, Mr. William Hui. There will be a Q&A session after the prepared remarks.

Before beginning, we'd like to remind you that discussions during this call contain forward-looking statements made under the safe harbor provision of U.S. Private Securities Litigation Reform Act of 1995. Such statements are subject to risks, uncertainties and factors that can cause actual results to differ materially from those contained in any such statements. Further information regarding such risks, uncertainties or factors is included in our filings with the U.S. Securities and Exchange Commission. We do not undertake any obligation to update any forward-looking statements as required under the relevant law.

During the call, we will be referring to certain non-GAAP financial measures and supplemental measures to review and assess our operating performance. These non-GAAP financial measures are not intended to be considered in isolation or as a substitute for the financial information prepared and presented in accordance with U.S. GAAP. For information about those non-GAAP financial measures and reconciliation to GAAP measures, please refer to our earnings press release.

I will now pass it to Ning for opening remarks.

Thank you all for joining us today. This past quarter presented a more challenging operating environment than we've seen in recent periods, driven primarily by heightened regulatory uncertainty and a more cautious credit backdrop. While these factors weighed on parts of our business, we moved quickly to adjust our risk posture and protect asset quality. I'm pleased to share that these actions have been effective. And at the same time, our Internet Insurance segment continued to deliver solid growth, reinforcing the resilience and diversification of our platform. As we look ahead, we remain focused on disciplined execution and positioning the company for the next-generation fintech with AI and blockchain.

As part of our ongoing transformation, we continue to advance our agentic AI capabilities to enhance process efficiency and strengthen unit economics. These innovations are helping us offset the margin pressure associated with rising credit risk. Our agentic platform, Magicube is already demonstrating meaningful impact, improving sales conversion, elevating risk controls and driving greater overall productivity.

With that, let me walk you through the key business highlights for the quarter. First, turning to our Financial Services segment. We facilitated RMB 20.2 billion in loan origination during this quarter, up 51% year-over-year. Our repeat borrowing rate remained at a record high of 77%, in line with last quarter and 16 percentage points higher than a year ago, while the number of our total borrowers decreased by 11% to 1.3 million compared to the same period last year due to the tightening of credit policies. Our total cumulative borrower base increased by 21% year-on-year to 14 million. We also continue to see healthy structural improvements across our borrower base. Average size for new loans from our lending platform rose from RMB 7,000 to RMB 10,100, driven by our ongoing shift towards higher credit quality customer segments and better credit predictability from repeat borrowers.

We expect this favorable mix trend to continue as we continue to trade up for better quality borrowers. Our agentic AI has delivered remarkable productivity boost in our operations. For marketing, our AI-driven marketing agent continued to deliver strong results. It enhanced the customer profiling accuracy and expanded the pool of identified high-intent users by 38% quarter-over-quarter. In addition, our proprietary AI agent now generates tailored responses across a wide range of customer inquiries, effectively reactivating dormant users and driving a 15% increase in their ATP engagement.

For customer service, our LLM-powered service robot continued to strengthen its performance with response accuracy rising from roughly 80% to over 92%. Meanwhile, the rate of inquiries requiring escalation to human agents declined by nearly 15% quarter-over-quarter.

For quality control and risk management, we continue to optimize our multi-model models. Fraud detection coverage increased from a weekly manual sampling of 450 cases to 5,800 by agentic AI, while accuracy improved to 91%.

Now let's turn to capital allocation. As of September 30, 2025, our total outstanding loan balance is RMB 34.2 billion, representing 10% quarter-to-quarter growth. Our funding costs rose by 55 basis points during the quarter, in line with the sector trend. We are now included in the wide list of nearly 30 compliant funding partners under the new regulatory framework, positioning us as one of the leading players in the market.

On asset quality and credit risk, we continue to see industry-wide pressure this quarter. Although we proactively tightened our credit policies, our risk indicators edged up in Q3. As of September 30, our 1- to 30-day delinquency rate stood at 2.7%, while the 31- to 60-day and 61- to 90-day delinquency rates were 1.7% and 1.4%, respectively. The good news is that we see the risk indicators for the loan portfolio from new borrowers begin to trend down in November, which is a proof of effectiveness of our upgraded credit strategy. However, from a conservative point of view, we expect the industry-wide impact on the overall asset quality to continue in the fourth quarter and the recovery is likely to begin early next year as the market stabilizes.

Our AI-driven collection capabilities played an important role in mitigating early-stage delinquencies. This automation drove productivity growth, reducing labor costs by an average of RMB 5 million per month, up from RMB 2.7 million in the second quarter, while improving service quality.

Turning to our overseas business. Our Indonesian operations launched on schedule in September 2025, and we expect this segment to contribute significant growth in 2026. Now turning to our insurance brokerage business. After navigating significant regulatory headwinds and commission pressure in 2024, we entered 2025 with a transformed operating model. Our insurance business has shifted from a high-touch, high-cost brokerage approach to a digital low customer acquisition cost, high-margin model by tapping into new insurance demand within our existing customer acquisition channels in the platform. This has allowed us to focus on a healthier, more profitable customer base that is contributing meaningfully to segment margins.

In the third quarter of 2025, gross written premium reached RMB 1.15 billion, an increase of 35% quarter-over-quarter. Revenue from the segment was RMB 84.2 million, up 45% from the prior quarter. Our Internet Insurance business continued its rapid expansion, delivering RMB 196 million in annualized premium, representing 204% quarter-over-quarter growth. Total customer number rose 93% quarter-over-quarter to 229,353, driven by more precise marketing and still low penetration within the target segment. We expect the Internet insurance business to sustain strong momentum over the coming quarters.

Finally, while we continue to strengthen and scale our core business, we are also investing strategically into the future. Building on our technology capabilities and our position within the broader fintech ecosystem, we are exploring new ways to better serve customers and manage assets through AI and blockchain-enabled solutions. We see AI and blockchain as core strategic pillars for the future of our business, especially as we expand our footprint globally. We are investing in the systems and capabilities needed to build our next-generation fintech infrastructure while deepening partnerships with key industry players.

In October, we signed an MOU with ChainUp, a leading crypto solutions provider in Singapore. And we also announced our plan to launch an Ethereum staking service, which is currently undergoing testing. This initiative marks an important milestone in our journey towards delivering seamless 24/7 global financial services. Over the next few quarters, we look forward to introducing additional products designed to enhance financing efficiency and asset monetization for our customers.

To conclude on the quarter, while the third quarter brought its share of challenges, the progress we've made demonstrates that our diversification and the forward-looking strategy are working. We've built a stronger, more resilient foundation that positions us well for sustainable growth and value creation in the quarters ahead. I'm confident that by staying disciplined and continuing to execute on our priorities, we will emerge even stronger.

With that, I will now pass it over to William, who will provide more details on the financials for the quarter.

Thank you, Ning. Hello, everyone. I will now walk you through our financial performance for the third quarter this year. Please refer to our earnings release and IR deck for further details, both available on our website.

For the third quarter, the total revenue grew by 5.1% year-over-year to RMB 1.55 billion, mainly attributable to 70% growth from the Financial Services segment. It was partially offset by the decline in revenues from the Consumers and Lifestyle segment as we announced to decommission the business in the fourth quarter of 2024.

In the Financial Services segment, total loan facilitation volume increased by 51% year-over-year to RMB 20.2 billion in the third quarter. The increase was driven by growth in average loan ticket size, the growth of repeated borrowers and increase in loan referral revenue. The loans from repeat borrower accounts for 77% of the total loan volume facilitated in the third quarter this year, up 16 percentage points compared to the same period last year. As the credit from the repeated borrower is more predictable, it allows us to extend the credit without substantially affecting our portfolio risk. The average size for new loan from our lending platform, Yixianghua, grew by 44% to RMB 10,100. Overall, the revenue from this segment increased by 70% year-over-year to RMB 1.4 billion in the third quarter. The revenue growth is driven by our loan guarantee services revenue, which reached RMB 458 million in the third quarter, up nearly 2.4x year-over-year, driven by higher loan facilitation under the risk-taking model.

As our service revenue and loan facilitation from the risk-taking model increases, our provisions for contingency liability also increased by 68.8% year-over-year to RMB 460 million. But as the economic benefits of the guarantee services is recognized over the next few quarters, a total of guarantee liabilities of RMB 930 million will be recognized as a revenue over the next few quarters. The contribution margin for the entire financial services segment improved from 5.2% in the third quarter of 2024 to 23% in the third quarter this year because of a higher revenue take rate and also a higher percentage of the deferred revenue from the guarantee business to be recognized as the revenue and also the higher borrower acquisition efficiency, which results in a 27.1% decrease in the origination expense, while the revenue grew by 70%.

In the Insurance segment, our gross written premium in the third quarter was RMB 1.15 billion, up 35% from the second quarter this year. It is showing a sign of recovery for this business. Compared to third quarter 2024, the premium is still down by 15%. The total revenue from the insurance line in the third quarter was RMB 84.2 million, up 44.9% quarter-on-quarter, but it is still down by 1.5% year-on-year. We have successfully turned around the business. The main growth contributor is the Internet Insurance line that we launched in the first quarter.

In the third quarter, the gross premium from the Internet Insurance line was RMB 196 million, and that represents 204% growth quarter-over-quarter. We expect this growth momentum will continue in the next few quarters and have significant revenue contribution to the overall insurance line. One thing to highlight is that the margin and the take rate for Internet Insurance business is much higher than the traditional brokerage line because the client for this segment comes from our existing customer traffic from insurance and other business segments. These customer segments are a better risk quality that traditional insurance carriers are not able to reach. As such, the Internet Insurance business has lower customer acquisition cost, better revenue sharing with the carriers and no commission cost. The margin is expected to increase as the premium scales, which will benefit the bottom line.

On the expense side, sales and marketing expenses in the third quarter decreased by 1.2% year-over-year to RMB 332 million. The marketing expenses decreased when our total loan facilitation increased by 51%. This is the result of the better AI-assisted precision marketing that drives a higher sales conversion, effectively lower the borrower acquisition cost. Research and development expenses decreased by 39% year-over-year to RMB 92 million. This is because during the same period last year, there was a one-off large system development project.

The origination, servicing and other operating costs decreased by 27% year-over-year to RMB 150 million because of the 27.1% decrease in the origination expense from the financial services business due to the improved collection efficiency driven by AI and lower commission costs from the traditional insurance brokerage line.

General and administrative expenses for the quarter increased by 30% year-over-year to RMB 104 million, primarily due to increased personnel-related costs to strengthen our risk management and to fund the plan for new business initiatives such as the development of the next-generation fintech that we mentioned in the announcement in October.

The allowance for contract assets and receivables and others for the quarter increased by 142% year-over-year to RMB 229 million. This is driven by higher receivables from loan facilitation service and guarantee services as the loan volume has grown with particularly strength from the risk-taking model that generates higher service revenues, along with the increase in the self-funded loan balance in the third quarter of 2025. Provisions for contingent liability this year increased by 69% year-over-year to RMB 460 million because of the increase in loan volume facilitated under the risk-taking model.

Net income for the third quarter was RMB 318 million, translating to RMB 3.65 per ADR shares or USD 0.51 per ADR shares. This represents 12% decline from the second quarter of this year. The pressure on profitability is attributed to multiple reasons, including the substantial upfront provisions under our risk-taking loan facilitation model, industry-wide volatility in asset quality, a declining fee rate for loan facilitation business following the new regulation as well as a decreasing commission rate in our traditional insurance brokerage line. Our net margin declined slightly from 22% in the prior quarter this year to 20%. However, we maintain a very good cash position. The net cash outflow from the operation in the third quarter was RMB 5.5 million, and our balance sheet remained robust with a total cash equivalent and restricted cash of RMB 4.3 billion. So this will position us well to address any future challenges and to capture new opportunities.

Looking ahead, we remain cautiously optimistic about our business, while we anticipate volatility in the credit and regulatory risk environment. Our disciplined credit policy, enhanced risk management capability and effective risk revenue model will position us well in this market environment. Our international business and Internet insurance segments are expected to drive a higher revenue growth and margin growth in the next few quarters. For the fourth quarter of 2025, we are projecting revenue to be in the range of RMB 1.4 billion to RMB 1.6 billion, so reflecting our disciplined approach to growth and risk management.

That's the end of my part of presentation. Thank you very much.

Thank you. And operator, we are open for Q&A.

[Operator Instructions]

The conference has now concluded. If you have any questions, you're welcome to contact the company's IR team. Thank you for attending today's presentation. You may now disconnect.

Thank you.

Hello, and welcome to the Deutsche Bank Virtual Investor Conference, dbVIC. This is Zafar Aziz from the Deutsche Bank team. I'm pleased to welcome our next presentation by Yiren Digital from China.

Before I introduce our speaker, a few points to note. Please click on the Questions box to ask a question. All of today's presentations will be recorded and can be accessed by the Deutsche Bank website, www.adr.db.com.

I'm happy now to hand over to Yiren Digital.

Hello, everyone. Thank you for joining Yiren Digital's second virtual roadshow this year by Deutsche Bank. My name is William Hui, and I'm the CFO of the company. We are the leading digital consumer lending and fintech platform operating in 3 major markets in Asia. We are building the next-generation fintech platform based on AI and Web3 technology.

Let me briefly walk through our history. So we began our digital lending business in 2006, and we are listed on the New York Exchange -- New York Stock Exchange in 2015. And in 2019, we expanded into insurance brokerage business through strategic restructuring and acquisitions. In 2024, we facilitated RMB 53.9 billion of loans or USD 7 billion to 6 million individual borrowers.

Since 2021, we have been investing heavily in AI R&D. We have our own GPU computing resource for model training. We built our own LLM and successfully integrated into our platform from -- on the process from a borrower acquisition, risk management and customer service. The technology innovation and industry know-how has prepared us to build our next-generation fintech.

So here are some of the investment highlights. We have strong cash flow from our lending and insurance business. They generated about USD 200 million of cash flow in 2024. The strong cash flow allow us to fund ongoing AI developments and Web3 initiative, which I will go through it later in the session, while maintaining financial strength.

We are expanding overseas to further diversify our business. We entered into Philippines in 2023 and Indonesia just 2 months ago. We expect the overseas business will contribute 10% of our business in 2026. On the technology side, this year, we introduced our legal and compliance LLM through a multi-AI agent platform, Magicube, which we will discuss later. Looking ahead, we will launch a Web3 platform to support our crypto assets and RWA-related services.

And finally, our share price does not reflect the potential of the technology mentioned above. Our market cap is roughly equal to our cash holding. So we will continue our dividend policy while investing in growth and maintaining a healthy cash flow.

So in second quarter of 2025, our financial services or the consumer lending accounted for 90% of our revenues. So in the second quarter, the financial services generated about RMB 1.5 billion or $214 million of revenue, representing 75% increase year-on-year. The growth is partially offset by the insurance brokerage business, which was facing an overall industry margin pressure. But in Q1, we started our online insurance business. The gross premium in the second quarter grew by 106% quarter-on-quarter. So that momentum is expected to continue in the second half.

Our online insurance business leverages existing lending traffic and apply AI-driven precision marketing. The online insurance segment generate at least 40% gross margins. Our AI capabilities are deeply embedded across our lending value chain. Our model analyzed 800 million data record accumulated over 19 years to classify borrower into 10 risk tiers. The system now reviews all users' submitted documents and detects about 30,000 suspicious entries a day. So that prevented about RMB 180 million of potential fraud losses last year.

On the marketing side, we work with 50 external traffic channels such as TikTok, Baidu, Tencent and Facebook for our overseas business. Our generative AI produces 1,700 personalized communication per day. So this improved the personalization, increased the repeated -- repeat borrowing rate from 65% in the fourth quarter of 2024 to 77% in the second quarter of 2025. On the loan recovery side, 81% of our first day delinquent loan is being handled directly by our AI-enabled recovery agents.

So these are the 2 AI innovations we have launched this year. So as mentioned earlier, our large language model, it is a multilingual sentiment-aware enterprise LLM supporting meeting analysis and communication planning and also the legal contract review, and it also automates the text file. The model has received a regulatory approval for commercial use, meaning we can deploy the model to external corporate customers.

Another innovation is the Magicube Agent Platform. This multi-model self-managed agentic AI platform integrates 6 specialized AI agents to handle critical and complex financial functions. It can generate text, voice and images. It supports customer service, compliance, sales, risk management and other business scenarios. One notable example is that the AI capital deployment agent is able to predict and deploy capital movement.

The 10-minute agentic AI process achieved 10 -- 100% capital deployment accuracy and generated a very thorough report, which would take 6 humans 1 week to do. One of the significant features of the Magicube is its self-maintenance features and the ability to manage multiple agents to handle complex tasks. We have monitoring agents that monitors the agent performance at 24 hours a day. If the performance of those agents begins to deteriorate, they will notify our R&D team for an upgrade.

The internationalization is one of our key strategies for our next-generation fintech. We expanded to Philippines in 2023 and achieved profitability after 14 months. Our Indonesian business just started the operation in September, and we expect to ramp up the volume in the first half of 2026. And we continue to explore other geographies for expansion.

The experiences in Philippines and Indonesia shows our technology platform and operation model is highly replicable for big market like Indonesia, which took us 6 months to build from zero to life, also in niche market like Philippines, where we can turn profitable within a short period of time.

So here's our financial highlights. In second quarter of 2025, our total revenue was RMB 1.65 billion or USD 236 million, representing 13% growth year-on-year. The growth is driven by 75% growth in the online lending business, partially offset by a 50% decrease in the insurance brokerage business. The overall revenue from the insurance brokerage business segment shows 3% increase quarter-by-quarter in Q2, helped by the triple-digit growth from the online insurance business.

On the net income side, we have reversed 5 quarters of net income decline and achieved a net income growth quarter-to-quarter in the second quarter of 2025. So driven by the AI automation that give us a cost saving and increased revenue.

For Q3, our business may be impacted by the new lending regulation and increase credit risk of the entire market, but the appreciation of the Ethereum, which we purchased in the first quarter, and the growth from the Internet insurance will help support the growth for the second half.

So now, we will discuss our future business. So in the future, we believe RWA tokenization and AI-enhanced financial operation will drive the next generation of fintech, and also, the next generation of fintech will expand our business internationally and use the RWA to provide monetization and price discovery for many different asset classes.

The RWA market is projected to reach $30 trillion by 2028. So our model spans the full value chain. So our business model encompasses the entire value chain. We begin with the coin custody services, providing custody for institutional clients and generate income through staking. We also offer exchange and collateralized lending, integrated with our existing lending businesses.

Once we have a good coin deposit base, we will launch asset management and hedge insurance to provide more diversified and value-added crypto-based investment products. If the traditional financial institutions are led by Goldman Sachs and Morgan Stanley and UBS and Deutsche Bank, so we hope to be one of the leading companies in the Web3 finance in the future.

So our first step is to build a blockchain technology platform. We established a technical development team in Hong Kong in April, and using the AI technologies, we will initially develop an Ethereum staking business and the infrastructure with a high cash flow efficiency and low regulatory risk. We will begin a small-scale functional release in early November with the Platform 1.0 to be launched in the first half of next year. So our target customers are the institutional clients who are holding cryptocurrency assets such as the funds and the listed company and the DAT company. So this is our strategy for the RWA.

The 3 major components of RWAs are assets, capital and platform. So regarding the assets, we have about $7.5 billion in assets that can be tokenized, including the U.S. treasuries, consumer finance loans and the AI algorithms and the computing power. So regarding the capital, we have access to numerous asset management channels overseas through our parent company CreditEase network. The platform connects assets, capital-enabling tokenization and Web2 and Web3 asset trading and decentralized finance.

So this is the ecosystem we plan to build. It will include not only our assets, products and channel, but also third-party assets, third-party capital and third-party developers. In early October, we signed a memorandum of understanding with a Singapore-based ChainUp to jointly develop a Web3 platform. The ChainUp provides a trading platform, which we provide the underlying assets. So we will collaborate to develop new products and technologies for market launch.

So here are our latest updates for our cryptocurrency business timetable. So as you see, we made our initial purchase of Ethereum in the first quarter. And in June of 2025, we established a digital asset task force in Hong Kong. And in 2025, we are establishing a partnership with different parties and also start building a beta systems, which will be launched later in November. And in 2026 first half, we will launch our 1.0 platform for staking service. And in the second half of this year, we will start rolling out our first fixed income product for the Web3 platform.

So for our valuation, as you see, Yiren Digital's share price has increased by 16% year-to-date. So it's outperforming other loan facilitation companies in China. So our dividend payout ratio is nearly 8%. It's already the second highest in the segment. So we have increased our dividends in September by 10% year-on-year, and we will continue to -- our dividend policy to reward our shareholders.

So our P/E ratio has not yet reflected the recent trends in Web3, and this is primarily due to the volatilities of the crypto market in October and coupled with the implementation of the new regulations for loan facilitation in China. So the investors have yet to adapt this changed environment, but we believe that there are a few Web3 listed companies with their own -- there are very few Web3 listed company that has a cash flow in the market.

So furthermore, driven by the internationalization and technical advancement of our business, there is a significant room for valuation growth. So this concludes my presentation part of the -- our road show. So I'm open for questions. Let me go through that.

Okay. So the first question, in a simple terms, how will you generate free cash flow from cryptocurrency in the future?

I would say, initially, when we built our staking business, it's a very -- it's relatively -- it's -- the development cost is relatively low, and we have targeted a few potential clients like some of the DAT company, which they have purchased a lot of Ethereum on their balance sheet, and they are looking to generate yields. So I think the ROI for the initial staking business is very high. So going forward, we are using those cash flow to invest in our next initiative, which is the crypto investment products and the other services that I just mentioned.

Okay. Next question. Can you elaborate on your approach to RWA tokenization and the opportunity size here?

So, so far, what we have done in our research is the -- in terms of the underlying assets for the RWA, we see there is -- private credit asset class has the highest growth so far. This is the coin-based statistics. So for the loans -- if you go back to our core business, which is the personal loan, which is considered a private credit as well. So we are building the -- building up the assets of which we already have. We lent about USD 7 billion of loan last year. And now, we are building up the other end of the equation, which is the capital, which we are working with a few asset management companies to do that.

So I think the key critical part is -- especially in Asia, is we are waiting for the issuance of the first stablecoins in Hong Kong. The legislation is in effect since August of this year. So I think the government expects to approve the first batch of a stablecoin in the first half of next year. So after that, we will see a big jump in terms of the adoptions of the stablecoin and the RWA.

Okay. Next question. How do you see the evolving regulatory environment in China affecting your consumer lending and the insurance brokerage business?

Okay. Let me speak for the consumer lending first. From the consumer lending in China, I would say it's -- that there is quite a lot of regulatory headwinds since April this year, as the government announced they will enforce the rate cap on the consumer lending. So that is in effect from October this year. So -- and so far, last weekend, I think the government made the announcement that they will rather relax on that regulation a little bit just to give the industry to better adopt the new environment. So that we see is positive. And -- but at the same time, we are preparing ourselves for this new regulation when the government decides to tighten the regulation again.

And on the insurance brokerage side, I think we have 2 parts of the business. One is the traditional insurance brokerage, which is quite affected by the commission fee -- commission rate cap. But on the other side -- hand is we -- our Internet insurance business has been growing very quickly. So it's been growing in the triple digit quarter-by-quarter.

So -- and in the second quarter, it accounts for about 20% of the revenue for our -- 20% of our insurance segment's revenues, and we see that to grow to more significant portions in the Q3 and Q4. So, so far from the -- on our insurance segment, it has definitely turned around as we see in our second quarter, our revenue -- we see a revenue growth compared to the first quarter. So I think our insurance business has turned around. And for the lending business, besides the regulatory headwinds, we are also monitoring the risk quite closely as well.

Okay. Let me go through the next question. How does blockchain integration fit into Yiren's long-term strategy? So they serve only as a treasury strategy?

We think our blockchain strategy is very strategic to us. So as we see the fintech is moving toward the Web3 era, so to speak, and we cannot miss out these opportunities. And we believe our business is ready for the blockchain integration because, first of all, we have a very strong AI that will help us to speed up the technology development.

And also, the -- our traditional know-how on lending because so far, when we talk to a lot of industry players, particularly on the Web3 space and -- or those Web3 native companies, they all have a very sophisticated trading platform and the infrastructure. But a lot of time, when they try to integrate their business to the traditional Web2 environment, I think they need some help. Like, for example, when I was talking to one of the crypto exchange in Hong Kong and that they are looking -- they want to do a Bitcoin collateralized lending. So that involves a risk management on the credits, which they have very little experience of. So they are looking for us to work together and provide those risk management know-how. So I think those are the angles that we see the opportunities.

Okay. Next question. Do you expect to continue generating positive net income margins? Or could the increase in regulation potentially lead to a further decline?

The way we see it is that we will continue to generate a positive income margins. No question about that. If you look at our 2024, our net income is USD 200 million, and it is still growing. And yes, the regulation will give us some pressures on the margins. But we believe there are different parts of our business that are growing very well. For example, the Internet insurance business that I just mentioned, and also, the -- our crypto business will start contributing revenues next year. So these are the -- I would say, the positive side and that sort of will offset some of the downside from the regulatory headwinds. But at the same time, we see the overall credit risk cycle have bottomed in the fourth quarter of this year. As we monitor the numbers in October and September, we're seeing it's trending -- that the risk is trending down. So that will be positive for us.

So okay. So I think we are running out of time. Thank you very much, and you can keep us -- you can -- we have the X account in the link, and so you can see our news updates from those websites. So thank you.