管理層發言
Ladies and gentlemen, thank you for standing by, and welcome to the Qifu Holdings First Quarter 2026 Earnings Conference Call. All participants are in a listen-only mode. There will be a presentation followed by a question-and-answer session. If you wish to ask a question, you will need to press the star key followed by the number 1 on your telephone keypad. Please also note today's event is being recorded. At this time, I would like to turn the conference over to Ms. Karen Ji, Senior Director of Capital Markets. Please go ahead, Karen.
Thank you, Dassie. Hello, everyone, and welcome to Qifu Holdings first Quarter 2026 Earnings Conference Call. Our earnings release was distributed earlier today and is available on our IR website. Joining me today are Mr. Haisheng Wu, our CEO; Mr. Alex Xu, our CFO; and Ms. Yan Zheng, our CRO. Now I will quickly cover the safe harbor statement. Today's discussions may contain forward-looking statements, particularly statements about our business and the financial results, that are subject to risks and uncertainties which could cause actual results to differ materially from those contained in the forward-looking statements. Please refer to the Safe Harbor statements in our earnings release, which also contains a reconciliation of the non-GAAP financial measures to GAAP financial measures. Now I will turn the call over to Mr. Haisheng Wu. Please go ahead.
Hello, everyone. Thank you for joining us today. Since April 2025, China's consumer credit industry has undergone profound structural adjustments under regulatory guidance. Entering Q1 this year, demand for consumer credit remained soft, and asset quality faced broad-based pressure. Household short-term consumer loan balances declined for the fifth consecutive quarter, decreasing by approximately RMB470 billion or 5% sequentially. In this challenging industry environment, we have upheld compliance, prudence, and high quality as the core principles of our operations. Rather than pursuing scale, we proactively optimized our user and asset mix to strengthen the overall health and long-term resilience of our business. Building on the proactive measures we implemented in the second half of last year to enhance risk management and business operations, we delivered a resilient performance in Q1, with notable improvements in risk indicators and operational efficiency. Our AI-powered credit decision engine: as of the end of Q1, our asset distribution platform served 167 financial institutions, delivering intelligent digital credit services to over 64 million credit line users on a cumulative basis. In Q1, we maintained rigorous risk standards against the backdrop of a softening retail credit market. As a result, total loan facilitation and origination volume on our platform declined by approximately 7.5% sequentially to RMB65 billion. Non-GAAP net income declined by 11.6% sequentially to approximately RMB950 million, while non-GAAP EPS on a fully diluted basis decreased by 6.4% to RMB7.7. Excluding one-off items, the take rate improved sequentially. In the second half of 2025, we continuously tightened risk policies, and this forward-looking strategy began to translate into tangible results in Q1. During the quarter, we further iterated and optimized our underlying risk capabilities across the entire credit life cycle. As a result, our FPD7, a leading risk indicator for new loans, declined by approximately 20% in Q1 compared with Q4 last year. As legacy loans continue to run off, portfolio-level risk metrics also improved month over month. By March, the C2M2 ratio, the risk indicator that measures the outstanding delinquency rate after 30 days of collection, returned to levels seen in July and August 2025. For the quarter as a whole, the C2M2 ratio decreased by roughly 17% sequentially to 0.8%, largely achieving our risk optimization targets. Specifically, these improvements were driven by the following initiatives. In the pre-loan and in-loan stages, we further strengthened our ability to identify high-quality customers while proactively screening out higher-risk segments. In the pre-loan stage, we upgraded the income and drawdown prediction models in our application scorecard, or A-scorecard, to more accurately assess user income and borrowing intent, which enabled us to serve more high-quality users. In the in-loan stage, we further refined our behavior scorecard, or B-scorecard, enabling targeted strategies such as credit line adjustments, rate reductions, and flexible repayment options for higher-quality borrowers. We also continuously updated our risk models to capture potential risk exposures. For example, when previously low-risk borrowers experience income fluctuations or take on multiple loans, our system could quickly detect these changes and proactively mitigate risk by reducing credit lines or raising approval thresholds. As a result of these efforts, average FPD7 for loans issued between January and March declined by approximately 5% compared to December last year, which provides a solid safety cushion against potential market volatility. In the post-loan stage, we continue to optimize our collection strategies. Since January, our day-1 delinquency rate has shown an overall downward trend, with the Q1 rate decreasing by roughly 7% sequentially, easing pressure on our collection front. Against this backdrop, we scaled back less cost-effective collection efforts and improved the efficiency of our resource allocation. At the same time, we upgraded the capabilities of our collection scorecard, or C-scorecard, by incorporating new features that reflect recent market conditions and shifts in user behavior. This enabled us to differentiate users more accurately by risk level and repayment willingness and to match each segment with the most appropriate collection approach. Through these efforts, we were able to manage risk while optimizing costs effectively, enhancing our collection efficiency. Together, these measures contributed to a steady month-over-month improvement in our 30-day collection rate during the quarter, with a quarterly average of 85.8%, up 1.8 percentage points sequentially. On the customer acquisition front, we maintained a disciplined approach, continuously optimizing acquisition channels and improving efficiency. In Q1, our overall acquisition costs fell by approximately 17% sequentially, with unit acquisition costs remaining largely stable compared to Q4. In parallel, we strategically increased marketing spending on high-quality users to further refine our user mix and build a pipeline of high-quality assets. In Q1, spending on this segment increased by approximately 40% sequentially. High-quality users tend to carry much lower risk than regular segments, with higher utilization, steadier long-term demand, and more repeat borrowing. This shift in our user mix will strengthen our portfolio quality and build a more resilient and sustainable moat for our business. Meanwhile, we substantially cut back on underperforming channels within the embedded finance model, helping to improve the risk and return profile of new users. On the funding front, the industry continued to face liquidity pressure during the quarter. By further increasing the proportion of ABS in our funding mix, we were able to reduce funding costs by approximately 10 basis points sequentially. In Q1, ABS issuance totaled RMB2.9 billion, up 16% from the prior quarter. For the remainder of the year, we will align the pace of our ABS issuance with on-balance-sheet loan origination to maximize capital efficiency. Since April, as industry adjustments continue, liquidity in the funding market has also tightened. To navigate periodic market volatility, we will continue to optimize our funding structure and diversify our partnerships with financial institutions to ensure sufficient funding supply in a volatile market while striving to keep our overall funding costs stable. Turning to our tech solutions business, we have continued to deepen collaboration with financial institutions and actively cultivate our enterprise-facing technology offerings as another long-term strategic pillar supporting banks in serving customer segments priced between 3% and 12%. At this stage, we are focused on validating these capabilities at scale, which will lay a solid foundation for long-term commercialization opportunities ahead. In Q1, loan volumes empowered by our tech solutions business reached RMB9.96 billion, representing 7-fold year-over-year growth. This demonstrates that our tech-driven capital-light model is steadily gaining industry recognition and being validated across multiple use cases. Our credit-focused AI agents have also entered initial commercial deployment. For example, one of our core AI agents, AI Loan Officer, is being deployed at a city commercial bank covering its retail, SME, and corporate business lines. With our FocusPro credit solution, we help banks serve small businesses and individual customers more efficiently by applying digital tools across their full credit life cycle from customer acquisition and risk profiling to day-to-day operations. This not only expands the scope of our business but also reflects our commitment to supporting the real economy and promoting financial inclusion. At the technology foundation level, we set a more ambitious long-term goal to fully transform the company into an AI-native organization. Central to this strategy is deep knowledge modeling. We are converting all historical documents, strategy libraries, and operational experience into structured context for large language models, creating a truly queryable knowledge base. With this foundation, departments across risk management, product, design, marketing, and engineering can integrate AI into their core operations. This not only provides assistance in day-to-day work but also fundamentally enhances the professional competence, decision-making quality, and professional boundaries of personnel across all departments. Cost savings and efficiency enhancements, often highlighted by the market, will be a natural byproduct of this evolution toward an AI-native organization. For example, AI coding tools have achieved impressive adoption across our engineering teams. As of the end of Q1, 98.4% of technical personnel were using AI tokens, with key roles consuming tens of millions of tokens per person per day. This indicates that AI adoption within our engineering has reached penetration levels comparable to those at top-tier internet companies in China. Token usage has also shown a clear correlation with productivity gains. Looking ahead, we will steadily extend this AI leverage to more business scenarios, accelerating our evolution into an AI-native organization. As we continue to strengthen our core domestic business, we are also accelerating overseas expansion while carefully managing risk along the way. In Q1, we successfully launched operations in new emerging markets and continued to fortify local teams and refine risk models in markets where we are already active. Leveraging our combined strengths in global capital, advanced technology, and local operational expertise, we aim to build a robust international presence, expanding efficiently and operating safely across multiple markets. Looking ahead, as the industry continues to adjust and restructure, we expect short-term uncertainties to persist. That said, the ongoing shakeout is creating a more structured and efficient market environment, offering a prime opportunity for industry leaders to strengthen and consolidate their positions. We remain committed to our one-core, two-wings strategy, with our domestic credit business as the core and tech solutions commercialization and overseas expansion as the two wings driving sustainable high-quality growth over the long term. Thank you. With that, I will now turn the call over to Alex.
Thank you, Haisheng. Good morning and good evening, everyone. Welcome to our first quarter earnings call. It was another quarter of challenging market conditions and tightening regulatory scrutiny, which caused further changes in the industry landscape and participants' behavior. We continue to focus on mitigating risks, improving efficiency, and reducing cost under such macro headwinds. Total net revenue for Q1 was RMB3.91 billion versus RMB4.09 billion in Q4 and RMB4.69 billion a year ago. Revenue from credit-driven services, capital-heavy, was RMB2.96 billion in Q1 compared to RMB3.43 billion in Q4 and RMB3.11 billion a year ago. The year-on-year decline was mainly due to a decrease in off-balance-sheet loan volume more than offsetting the increase in on-balance-sheet loans. The sequential decline was due to lower overall capital-heavy loans as well as a decrease in average pricing of the loans. Overall funding cost declined roughly 10 basis points quarter-on-quarter as we further optimized the funding mix with increased percentage contribution from ABS in Q1. Revenue from platform services, capital-light, was RMB951.9 million in Q1, compared to RMB660 million in Q4 and RMB1.58 billion a year ago. The year-on-year decline was mainly due to significantly lower ICE contribution in response to the regulatory changes. The sequential increase was mainly due to a better ICE take rate due to improved risk. During the quarter, average IRR of the loans we originated and/or facilitated was 18.7%, compared to 19.5% in the prior quarter. As we continue to focus on extracting high-quality users, in the coming quarter we may see modest fluctuation in average pricing under the current regulatory framework. Sales and marketing expenses declined 17% quarter-on-quarter and 23% year-on-year. We added approximately 1.19 million new credit line users in Q1 versus 1.45 million in Q4. We took a more cautious view in customer acquisition and will continue to maintain a controlled pace to acquire new users in the near term in response to the changing regulatory direction and still uncertain macro conditions. The 90-day delinquency rate was 3.5% in Q1 compared to 2.71% in Q4, which reflects the elevated risk level near the end of 2025. Again, 90-day delinquency is a lagging indicator and internally it is not a metric we focus on as much. Day-1 delinquency rate was 5.7% in Q1 versus 6.1% in Q4. 30-day collection rate, C2M2, which represents the outstanding delinquency rate after 30 days of collection, was 85.8% in Q1 versus 84.1% in Q4. After 30 days of collection, the outstanding delinquency rate was 0.8% in Q1 versus 0.97% in Q4. The noticeable risk improvement in Q1 was mainly related to our risk-tightening measures and loan mix shift toward new loans, and we saw continued modest improvement in overall risk performance in recent months. Given current macro conditions and the regulatory changes, we continue to take a prudent approach in booking provisions against potential credit losses. Total new provisions for risk-bearing loans in Q1 were approximately RMB1.68 billion versus RMB1.92 billion in Q4, mainly due to lower loan volume, partially offset by an increased provision booking ratio. Write-backs of previous provisions were RMB308 million in Q1 versus RMB274 million in Q4. Despite improved new loan risk quarter-on-quarter, provision coverage ratio, which is defined as total outstanding provision divided by total outstanding delinquent risk-bearing loan balance between 90 and 180 days, was 391% in Q1 compared to 481% in Q4. The temporary decline in provision coverage ratio was mainly due to a higher risk level in Q4 driving up the delinquent risk-bearing loan balance between 90 and 180 days in Q1, while total outstanding provision was largely unchanged quarter-on-quarter. Non-GAAP net profit was RMB946 million in Q1 compared to RMB1.07 billion in Q4 and RMB1.93 billion a year ago. Significant year-on-year decline in profitability was mainly due to lower loan volume and lower pricing, higher credit cost, and deleveraging in operations. Non-GAAP net income per fully diluted ADS was RMB7.7 in Q1 compared to RMB8.23 in Q4. Effective tax rate for Q1 was 21.6% compared to our typical ETR of approximately 15%. The higher-than-normal ETR was mainly due to withholding tax on dividend distribution from onshore to offshore, and a more prudent tax provision under heightened tax regulatory scrutiny. Leverage ratio, which is defined as risk-bearing loan balance divided by shareholders' equity, was 2.4x in Q1 versus 2.7x in Q4 due to lower risk-bearing loan balance. We expect to see the leverage ratio fluctuate around this level in the near future. We generated approximately RMB2.1 billion in cash from operations in Q1 compared to RMB3.15 billion in Q4. Total cash, cash equivalents, and short-term investment was RMB10.79 billion in Q1 compared to RMB10.72 billion in Q4. In Q1, we continued to buy back our outstanding convertible bonds. As of May 26, 2026, the company had repurchased approximately US$577 million in aggregate principal amount of the CBs for US$502 million in cash on the open market and in off-market privately negotiated transactions. Approximately US$113 million in aggregate principal amount of the CB remained outstanding. The repurchase of the CB allowed us to reduce our long-term debt obligation and associated interest payments at favorable terms, potentially strengthening our financial position and flexibility. We will continue to optimize our capital allocation strategy to reflect the changing macro dynamics to support business initiatives and to return capital to shareholders. As we maintain a progressive dividend policy, we may start to opportunistically look into an entry point to resume share repurchase even though the macro and regulatory environment is still unsettled. Finally, regarding our business outlook, while we are encouraged by the noticeable improvement in risk metrics, macro uncertainty and regulatory pressure will likely persist in the near future. We will continue to take a cautious approach in business planning for 2026 and focus on risk control, efficiency improvement, and cost cutting. For the second quarter of 2026, the company expects to generate non-GAAP net income between RMB900 million and RMB980 million, representing a year-on-year decline of between 47% and 51%. This outlook reflects the company's current and preliminary view and is subject to material changes. With that, I would like to conclude our prepared remarks. Operator, we can now take some questions.
分析師問答
Thank you. Please press star then 1 on your telephone, and wait for your name to be announced. If you wish to cancel your request, please press 2. If you are using a speakerphone, please pick up a handset to ask your question. For those who can speak Chinese, please start your question in Chinese followed by English translation. To allow enough time to address everyone on the call, please keep it to one question and one follow-up, and return to the queue if you have further questions.
Yeah. I have two questions. Just observing the average pricing of the loans, it continued a downward trend relative to relatively stable pricing at peers. What was the rationale and thinking behind that? Also, where will the average loan pricing eventually settle later this year? And based on the current business model, what is the average demand as well as business scale going forward and how will that impact shareholder returns? Thank you.
Okay. Thank you, Richard. Let me take your first question and Alex may take your second one. In terms of pricing, first, as we are currently in a rate-cut cycle, lower financing cost helps improve households' balance sheets and unlock healthy consumption demand. As we mentioned before, this is overall positive for the industry as it is accelerating market consolidation. Long-term platforms that relied on high prices to cover high risk and aggressively grow their market shares in the past will eventually exit the market. At the same time, this sets a higher bar for market participants. Over the longer term, companies with more precise user profiling and better risk-based pricing capabilities are likely to take market share. With this in mind, we proactively adopted targeted pricing to optimize our user mix. Since Q4 last year, we have put more effort into acquiring high-quality users. This quarter, our spending on that segment was up by about 40% sequentially. As a result, the share of high-quality users in our new customer loan volume jumped 25 percentage points from Q3. At the same time, we also optimized pricing for our existing users with better risk profiles. By giving them more competitive offers, we intended to increase their stickiness on our platform. In addition, we also expanded their borrowing capabilities reasonably to help build a stable long-term LTV. As a result of these efforts, our average pricing was down 80 basis points sequentially. Based on our observation so far, high-quality users tend to carry lower risk than regular segments, with higher utilization, more repeat borrowing, and steadier long-term demand, and therefore much healthier LTV. Going forward, we will continue iterating our models to better identify higher-quality users, improve risk performance, and use flexible pricing strategies together with a great user experience to drive retention and repeat borrowing. All of this will further improve LTV for this segment. Strengthening our capabilities to serve high-quality users is a long-term play. In the short term, it requires investment, but in substance, it is a trade-off between near-term profit and long-term sustainable value. By building these capabilities we are also reshaping our business model, making it healthier and better positioned to navigate a more complex and fast-changing market environment. On the outlook for pricing, we increased our investment in acquisition and conversion of high-quality users starting in Q4 last year. Over the past two quarters, we have seen a step-change increase in the share of high-quality customers in total loan volume, with pricing trends moving in the same direction. As a result, our pricing has declined by 2.2 percentage points over this period. Going forward, we expect to maintain a relatively steady pace on acquisition and offering strategy. As the user life cycle continues to evolve, our mix will gradually shift to higher-quality users. Meanwhile, we will continue to optimize pricing strategy for high-quality and regular users so as to strike a balance between risk and profit. Over the longer term, we will remain flexible and adjust our pricing strategy based on regulatory and market changes. At the same time, we will continue to improve our risk models and operational efficiency to strengthen profitability. Okay, Richard.
I will take the second part of your question. The landscape of the consumer finance industry is still changing and the near-term macro and regulatory environments remain complicated. On our side, we are actively adjusting our asset mix and exploring ways to diversify our business. Market demand is always there, and as we keep upgrading our capabilities in serving high-quality users, we are confident in our ability to meet more credit needs over time. However, at this stage, our focus probably remains on the health and sustainability of the business as opposed to scale. In terms of shareholder returns, our balance sheet is still solid and we have a strong capital base to support both business growth and shareholder returns. At the same time, our business continues to generate substantial profit and healthy positive cash flow which steadily builds up our capital base. Both dividend and buybacks are options for us. At this stage, we see dividend as a way to give shareholders certainty in an uncertain environment, and assuming a stable regulatory environment we will maintain a progressive EPS policy by flexibly adjusting the dividend payout ratio. On the buyback side, we believe our current valuation is attractive and our net assets far exceed our market cap. We have strong conviction in our intrinsic value. Although there is still uncertainty around the macro and regulatory environment and our business model continues to evolve, we believe share buyback once again becomes a viable option for shareholder returns at this junction. Thank you.
Thank you. Your next question comes from Alex Ye with UBS. Please go ahead. Pardon me, Alex. Your line may be muted.
So my question is regarding the loan growth outlook. Given the stabilization of asset quality and that you have already shifted your customer mix toward higher-quality customers, when could we expect the company to increase its risk appetite a little bit? Should we expect loan volume to return to some sequential growth in the coming quarters?
Thank you. First, the structural changes we have made to risk management and our user mix have delivered clear results since the second half of 2025. We have implemented a series of risk-testing measures, significantly reallocating resources toward the acquisition and engagement of high-quality users. These initiatives have led to a marked improvement in both our asset structure and asset quality. In Q1, our risk metrics improved month by month; the C2M2 ratio in March returned to levels seen in July and August of last year. In April and May, the C2M2 ratio stayed stable and continued to improve marginally. While keeping risk stable and improving, we have also been exploring structural growth opportunities. For example, for customers with a solid safety margin based on our risk model, we dynamically adjusted our approval and credit line rules. Within a safe range, we improved our conversion rate and effectively served more credit demand. We have also optimized our asset distribution strategy and worked with funding partners to explore innovative compliant business models to improve acceptance rates at the partners. In addition, we have expanded our product offerings to serve long-tail customers throughout their life cycle, improving our ability to serve them while maintaining a healthy risk profile. That said, there is still some uncertainty around the regulatory environment and near-term adjustments are still ongoing. Therefore, we will continue to maintain a prudent approach in our overall business strategy. At this stage, our focus is on improving our user mix, asset health, and operational efficiency. We will not chase volume growth blindly, and we will not rush to loosen risk standards just because of short-term improvement in asset quality. By continuously improving our asset mix and the overall platform resilience, we will be better positioned to navigate potential market volatilities in the future. Thank you.
Thank you. Your next question comes from Cindy Wang with China Renaissance. Please go ahead.
Thanks for taking my questions. I have two questions. First, could you give us some color on domestic risk performance in April and May? Has overall credit risk continued to improve from Q1? Second, what is the trending customer acquisition cost? What customer-acquisition channels are currently being used to acquire high-quality customers, and what is the risk performance from new customers? Thank you.
Okay. Yan Zheng, can you take the first one? I will take the second one.
Thank you. Okay.
Let me briefly translate Ms. Zheng's answers. Our risk metrics improved meaningfully in Q1, with the C2M2 ratio down 17% sequentially to 0.8%. By March, the ratio had returned to the levels seen in July and August last year. Overall, the improvement was greater than we had expected and we have largely achieved our risk optimization targets. This was mainly driven by our continuous optimization of risk strategies and iterations of our models. First, we started tightening risk standards since the second half of last year at the underlying capability level. We kept upgrading our A- and B-scorecards to better separate high-quality users from high-risk ones. As a result, the FPD30 metric for new loans has continued to improve. FPD30+ decreased by 18% in Q4 compared to Q3, and further decreased by approximately 22% in Q1 compared to Q4. Moreover, based on early performance of new loans issued in April, FPD7 was roughly flat versus March, continuing to stay at a desirable level. Meanwhile, our average loan tenure is around 10 to 11 months. As new loans with lower risks take a larger share and legacy loans gradually roll off, our asset quality will improve eventually, and this cycle usually takes about two to three quarters. Given this timeline, our turnaround in Q1 asset quality was fully expected. Secondly, on the collection side, we also upgraded our C-scorecard over the past two quarters. We added new features that capture recent market conditions and changes in user behavior. Based on risk tier, we also match collection strategies more precisely, including better user reach-out and more tailored prepayment solutions. This data-driven approach has worked very well. Our collection rate improved month by month, with Q1 collection rate increasing by 1.8 percentage points sequentially. Going forward, we will continue to refine our strategies to make collection management more efficient while keeping collection rates steady. Based on the risk performance we observed in April and May, we expect the C2M2 ratio to remain generally stable with a positive trend compared to March. If everything proceeds normally in June, we anticipate further risk improvement for Q2 compared to Q1. However, regulatory uncertainty remains and the industry will also face some adjustments as certain policies take effect. Therefore, we will stay cautious on risk policy for now and keep optimizing marginal assets and acquisition channels to build some buffer against potential market headwinds going forward. Thank you.
In terms of customer acquisition, our overall strategy this year is to spend cautiously as we aim for structural growth and constantly improve marketing efficiency. In Q1, we kept prudent marketing spending while reallocating resources. We cut back our spending on regular segments and focused on higher-quality users. The upfront acquisition cost for high-quality users is typically higher, but thanks to our improved marketing efficiency, we kept our blended acquisition cost roughly flat sequentially while further improving our user mix. In Q1, our spending on higher-quality users was up about 40% while our spending on regular segments came down significantly. We leveraged similar channels such as feed-based channels to acquire high-quality users while using upgraded acquisition models to help us precisely identify users with strong credit profiles and stable repayment ability. Meanwhile, we have higher tolerance for bidding prices for acquiring these customers. As a result, the share of high-quality users among new credit line users was up 6% sequentially. High-quality users not only have better early risk performance, they also show more stable operating metrics later on like more stable repeat borrowing rates and higher balance retention. That said, we are still in an early stage of building our know-how to serve this segment. Going forward, we will keep iterating our risk models based on our own user behavior and use more refined operations to steadily improve their lifetime value. We will stick to this acquisition strategy by focusing on high-quality users and keep optimizing marginal channels and assets. I want to emphasize that customer acquisition cost is just a number, which is heavily impacted by channels and acquisition mix. Therefore, we do not simply chase a low absolute CAC. Instead, for every dollar we spend, we track payback period, user LTV, and measure efficiency based on ROI. For API channels, the acquisition cost is more tied to loan volume, so we measure IRR on each loan to ensure every single loan is profitable. At the end of the day, we want to make sure every dollar we spend delivers solid returns.
Thank you. Your next question comes from Yada Li with CICC.
Thanks for taking my question. I have two quick questions. First, do we have further room to cut funding costs? Second, any update on our overseas business? Thank you.
I will take the first part. Regarding funding cost, even though the industry is still undergoing significant changes and we do face some liquidity pressure, we are optimizing our funding mix by maintaining a high percentage contribution from ABS funding. In Q1, ABS issuance reached RMB2.9 billion, up 16% sequentially. The portion of ABS in our external funding went up 6 percentage points sequentially. With a more optimized funding mix, we were able to reduce our funding cost by about 10 basis points sequentially. Looking ahead, in the coming quarters, we will continue to seek ABS issuance windows and align the pace with on-balance-sheet loan origination demand. We aim to maintain sufficient funding supply in the volatile market while keeping overall funding cost stable. Regarding operating cost, since the end of 2025, we have been improving operating efficiency across all functions, including front-, middle-, and back-office operations. On the direct cost side, our core approach is to dynamically manage operating resources, align them with business needs, and continue to optimize the cost structure. Take collection cost as an example: our portfolio risk metrics improved throughout Q1, gradually easing pressure on the collection side. After seeing this trend, we quickly scaled back high-cost, low-marginal-yield strategies and adjusted our collection resource mix, further improving our overall collection efficiency. As for customer acquisition cost, Haisheng already covered the details. On the G&A side, we took active steps in Q1 to improve efficiency mainly around optimizing redundant or inefficient roles. This helped us make our teams leaner and execution more efficient. The positive financial impact of these initiatives will be gradually reflected in the coming quarters. We are also using AI to upgrade our organization, as Haisheng mentioned earlier. Right now, we are feeding our past data, strategies, and experience into AI models to build a queryable knowledge base to support all our teams. This goes beyond just helping with day-to-day tasks; it helps our people improve skills, decision quality, and professional capability across the board. As we use AI more, the cost efficiency that the market cares about will come naturally as a byproduct as we evolve to a truly AI-native company.
In terms of global markets, overseas expansion is a key part of our one-core, two-wings strategy. This year, we will allocate more resources to the overseas business and accelerate the pace of our international presence. Based on extensive research into global markets, we have identified several target markets with great potential in terms of credit demand. In these markets, our data-driven risk management capability will make us stand out. In markets we have already entered, such as the UK and a Latin American country launched in Q1 this year, our current focus is on localizing risk models by gradually adding local credit data and open banking data to improve our risk-based pricing. Early results are broadly in line with our expectations. For other high-potential regions such as Southeast Asia, we are actively planning the next steps. Overall, our international business is still in the early stage of investment and capability building, but we are patient about the long-term opportunity. We will leverage global capital, cutting-edge technology, and local expertise to drive substantial long-term growth in our overseas markets over the next several years while keeping risk under control. We are confident that with solid execution over the next three to five years, we will become a truly global fintech company. Thank you.
Thank you. There are no further questions at this time. I will now hand back to management for closing remarks.
Okay. Thanks again to everyone who joined us today. This concludes our conference call. If you do have any additional follow-up questions, please contact us offline. Thank you.
Thank you. That does conclude our conference for today. Thank you for participating. You may now disconnect.