Hello, ladies and gentlemen. Thank you for standing by for KE Holdings Inc.’s second quarter 2025 earnings conference call. Please note that today’s call, including the management’s prepared remarks and question-and-answer session, will all be in English. Simultaneous interpretation in Chinese is available on a separate line for the duration of the call. To access the call in Chinese, you will need to dial into the Chinese language line. At this time, all participants are in listen-only mode. Today’s conference call is being recorded. I will now turn the call over to your host, Ms. Siting Li, IR Director of the Company. Please go ahead, Siting.

Good evening and good morning, everyone! Welcome to KE Holdings Inc., or Beike’s, second quarter 2025 earnings conference call. The Company’s financial and operating results were published in the press release earlier today and are posted on the Company’s IR website: investors.ke.com. On today’s call, we have Mr. Stanley Peng, our co-founder, chairman and chief executive officer, and Mr. Tao Xu, our executive director and chief financial officer. Mr. Xu will provide an overview of our business updates and financial performance, then Mr. Peng will share more strategic thinking on our current and future developments. Before we continue, I refer you to our safe harbor statement in our earnings press release, which applies to this call as we will make forward-looking statements. Please also note that Beike’s earnings press release and this conference call include discussions of unaudited GAAP financial information as well as unaudited non-GAAP financial measures. Please refer to the Company’s press release, which contains a reconciliation of the unaudited non-GAAP measures to comparable GAAP measures. Lastly, unless otherwise stated, all figures mentioned during this conference call are in RMB. Certain statistical and other information relating to the industry in which the Company is engaged to be mentioned in this call has been obtained from various publicly available official or unofficial sources. Neither the Company nor any of its representatives has independently verified such data, which may involve a number of assumptions and limitations, and you are cautioned not to give undue weight to such information and estimates. For today’s call, management will use English as the main language. Please note that the Chinese translation is for convenience purposes only. In the case of any discrepancy, management’s statements in their original language will prevail. With that, I will now turn the call over to our CFO, Mr. Tao Xu. Please go ahead.

Thank you, Siting, and thank you everyone for joining our 2025 half-year results conference call. In Q1, the real estate market continued the recovery momentum we saw at the end of last year. However, as we entered Q2, that momentum softened. The slowdown was largely due to international trade friction and the fading impact of earlier policy measures. Because of the high base created by the intensive easing policies in the middle of last year, the real estate market recorded a year-over-year decline in Q2. Turning to our business performance. Our platform’s agent and store network continued to scale, along with refined operations and ecosystem improvements. Our existing home and new home businesses significantly outperformed the market in the first half of 2025. The proportion of housing transactions from existing home sales reached a record high. At the same time, our home renovation and furnishing business and home rental services business both achieved high-quality growth. Revenue from non-housing transaction services accounted for 41% of total revenues in Q2, highlighting our diversified growth drivers. Regarding our overall financial performance in Q2, our total GTV was RMB878.7 billion, representing a year-over-year increase of 4.7%. Total revenues reached RMB26 billion, up 11.3% year-over-year. Gross margin declined by 6 percentage points year-over-year to 21.9%. GAAP net income was RMB1.31 billion, falling 31.2% year-over-year. Non-GAAP net income reached RMB1.82 billion, down 32.4% year-over-year. Next, I’d like to elaborate on the operational updates and financial performance of our business segments. Looking at our housing transaction services, the momentum from our proactive growth efforts has been clearly evident. In the first half of the year, the number of existing home sales transactions on our platform rose by 26%, outpacing the market’s growth rate of 19% estimated by the Beike Research Institute, both year-over-year. New home orders on our platform increased by 19%, outperforming the market which, according to Beike Research Institute, declined by 6%, both year-over-year. The share of our existing home sales continued to rise, with their proportion of total home transaction orders on the platform increasing from 51% in the first half of 2021 to 76% in the same period of this year. Our competitive edge in the existing home market led to more stable and solid overall business performance. Our agent and store network further expanded. In the first half of the year, various high-quality industry brands joined our platform, including Guangzhou’s Hope Real Properties, Tianjin’s Baoyuan Property, Wuxi’s Zhongshan Real Estate Service and Shenyang’s Yumei Property. The number of active stores on our platform increased by 30% in the first half of the year, with active non-Lianjia stores soaring by 36.8%, both year-over-year. The number of active agents on our platform leaped by 19.5% in the first half of the year, including nearly 24% increase in the number of non-Lianjia agents, both year-over-year. For our existing home transaction services, we continued to deepen and refine our operations, leveraging our scientific management system, platform-based operations, and AI-driven technological applications to boost store and agent productivity, throughout home listing, customer acquisition, and conversion process. On the home listings side, we have tools like home maintenance scores and ‘exceptional home product’ to help agents better market home listings, while focusing on top listings to increase transaction conversion. On conversion side, we also implemented measures to strengthen store level operations and network cooperation, to enhance matching and transaction efficiency. These measures included mechanisms like: incorporating competition into our stores’ scientific management system, deepening the operation of co-governance councils, store points-based incentive program, and designating quality-and-efficiency business districts. In terms of financial performance, revenue from existing home transactions reached RMB6.7 billion in Q2, down 8.4% year-over-year and remaining relatively flat quarter-over-quarter. GTV was RMB583.5 billion, remaining relatively stable year-over-year and quarter-over-quarter. The GTV growth outpaced revenue on a year-over-year basis, mainly due to a higher GTV contribution from existing home transaction services facilitated by connected agents, for which revenues are recorded on a net basis. The contribution margin for existing home transaction services was 39.9% in Q2, a decline of 7.5 percentage points year-over-year, primarily due to higher fixed labor costs resulting from an increase in the number of Lianjia agents and the lasting impact of the agent welfare improvement strategy we implemented last year. Sequentially, the contribution margin grew by 1.8 percentage points due to stronger leverage as our series of cost reductions and efficiency enhancements initiatives conducted this year led to a quarter-over-quarter decrease in fixed labor costs while revenue remained generally flat. For our new home transaction services, the scale of our collaborative projects remained steady. Through our AI-driven agent, Qianji, we refined the management of new home projects, revitalizing more existing projects. On the customer front, we reinforced business synergies between existing, new home and rental services, and fine-tuned operations, while launching our AI assistant, Qianzhi, to help service providers stimulate customer demand and improve matching efficiency. In terms of financial performance, our new home GTV reached RMB255.4 billion in Q2, up 8.5% year-over-year and 10% quarter-over-quarter. Revenue from new home transactions was RMB8.6 billion in Q2, rising by 8.6% year-over-year and 6.7% quarter-over-quarter. Revenue growth was in line with GTV growth year-over-year, demonstrating our steady monetization capabilities in new home transactions. While GTV growth outpaced revenue growth sequentially due to the seasonal fluctuation of the take rate. The contribution margin from new home transaction services fell by 0.6 percentage points year-over-year to 24.4%, due to an increase in variable costs resulting from our agent welfare improvement last year. Sequentially, the new home contribution margin rose by 1 percentage point, largely attributed to the quarter-over-quarter decline in variable costs thanks to our refined operations and focused sales strategies to maximize unit sales per property project this year. For our home renovation and furnishing services, we focused on enhancing the operations to build up our underlying capabilities to support our sustainable growth. On product capabilities, we analyzed customer data and insights to understand their core needs. Leveraging master designs and R&D, we introduced home renovation modules that can be flexibly configured and quickly iterated. By combining these modules with designer adjustments, we provide customers with a one-stop home renovation solution. On the supply chain side, our digital infrastructure enabled us to significantly streamline partner brand selection and SKU counts based on customer needs. The proportion of centralized procurement rose markedly, while the overall unit purchase price declined significantly. To further improve our delivery quality, we identified over 2,600 high-quality project managers on our platform, improving their efficiency and income, while elevating the end-user experience. Operationally, we implemented the ‘quality-and-efficiency business districts’ strategy. We scored business districts based on multiple indicators such as building age and housing transaction volume. This allowed designers, project managers, and other service providers to focus on high-scoring districts, enabling them to gain deeper insights into customer needs and property conditions. By introducing upfront site measurements and other processes, we reshaped workflows to improve operational efficiency. We also accumulated and refined design solutions and construction guidelines, ultimately enhancing the customer experience. In terms of financial performance, revenue from our home renovation and furnishing business reached RMB4.6 billion, increasing by 13% year-over-year. This was mainly driven by the increase in home renovation orders, alongside a higher average revenue per order, stemming from an increase in the average price of furniture and home furnishing retail. Contribution margin for the home renovation and furnishing business reached 32.1%, up 0.8 percentage points year-over-year, primarily driven by a larger proportion of centralized procurement, and enhanced order dispatching efficiency. Sequentially, the contribution margin fell by 0.4 percentage points, mainly attributable to a structural shift, with an increased revenue contribution of furniture and home furnishing retail, which has a relatively low contribution margin. In our home rental services business, we continued to iterate our products and applied AI to reconstruct our business processes and operational funnels. On the product front, we expanded our differentiated product portfolio, launching “product zero-nine” in the first half of the year to meet homeowners’ various needs around vacancy periods and rental income, while balancing risk and returns for our business. For unit signups and occupancy, we implemented quality-driven leads allocation rules, so that better listings and better service providers get more leads, and customers will also benefit. We also leveraged AI capabilities for intelligent collection and identification of rental housing conditions, as well as intelligent pricing, to explore a unit-signup model led by the platform's AI. In terms of operational management, we leveraged AI’s massive computing power to optimize resource dispatching, inventory, unit signups and occupancy, which enhanced both personnel productivity and rental occupancy rates in pilot regions. Our overall personnel productivity improved remarkably in the first half of the year. The average number of rental units managed per property manager rose significantly year over year, with the number of unit signups growing by over 50% in June compared with the same period of last year. Regarding financial performance, revenue from our home rental services business reached a record high of RMB5.7 billion in Q2, up 78% year-over-year, mainly benefiting from the rapid growth in the number of rental units under management. By the end of Q2, we had over 590 thousand rental units under our management, compared with over 310 thousand in the same period of 2024. The contribution margin for home rental services was 8.4%, up 2.5 percentage points year-over-year and 1.6 percentage points quarter-over-quarter, largely due to the improved gross profit of our Carefree Rent business. As we continued to refine the Carefree Rent business model, based on the essence of the service contracts, the revenues of some newly managed rental units were recorded as net revenues derived from service fees. For the Beihaojia business, our strategic direction is very clear and firm. We will never be developers. Our commitment to an asset-light business model is absolute. Other than the Chengdu Beichen and Shanghai Fengxian Xincheng project, we will not independently operate any other projects. In Q2, our revenue from emerging and other services decreased by 50.6% year-over-year and grew by 23.5% quarter-over-quarter to RMB432 million. Now moving to other costs and expenses, profitability, cash flow, and other financial metrics in Q2. Our store costs reached RMB762 million, increasing by 11.9% year-over-year and remaining relatively stable quarter-over-quarter. The year-over-year growth was mainly from higher store repair and maintenance costs. Other costs were RMB588 million, up 15.2% year-over-year and 7.5% sequentially, primarily due to higher basic maintenance costs of our home rental services business. Gross profit dropped by 12.5% year-over-year to RMB5.7 billion. Gross margin was 21.9%, down 6 percentage points year-over-year, primarily due to the decrease in contribution margin from existing home transaction services. Gross margin increased by 1.2 percentage points sequentially in Q2, mainly due to the greater revenue contribution from the home renovation and furnishing business, which has a relatively high contribution margin. In Q2, our GAAP operating expenses totaled RMB4.6 billion, up 3.1% year-over-year and 9.7% sequentially. Notably, G&A expenses were RMB2.1 billion, remaining flat year-over-year and increasing by 11% quarter-over-quarter, primarily attributable to an increase in bad debt provisions. Sales and marketing expenses amounted to RMB1.9 billion, remaining relatively stable year-over-year and growing by 7.1% quarter-over-quarter, primarily resulting from increased sales and marketing expenses for the home renovation and furnishing business. Our R&D expenses were RMB633 million, up 25.6% year-over-year and 8.5% sequentially, largely driven by higher personnel expenses and technical service fees. In terms of profitability, GAAP income from operations totaled RMB1.06 billion in Q2, down 47.4% from the same period of last year, and up 79.4% sequentially. GAAP operating margin was 4.1%, dropping by 4.5 percentage points from Q2 2024 and rising by 1.5 percentage points quarter-over-quarter. Non-GAAP income from operations totaled RMB1.61 billion, falling by 42.9% from the same period of last year and increasing by 40% sequentially. Non-GAAP operating margin reached 6.2%, down 5.9 percentage points from Q2 2024, mainly due to year-over-year gross margin decline. Non-GAAP operating margin rose by 1.3 percentage points from the previous quarter, mainly attributed to sequential gross margin improvement. GAAP net income totaled RMB1.31 billion in Q2, down 31.2% year-over-year and up 52.8% quarter-over-quarter. Non-GAAP net income was RMB1.82 billion, falling 32.4% year-over-year and increasing 30.7% quarter-over-quarter. Moving to our cash flow and balance sheet, we generated a net operating cash inflow of RMB826 million in Q2. New home DSO reached 51 days in Q2, remaining at a healthy level. On top of spending approximately US$254 million for share repurchases and distributing US$400 million for 2024 final cash dividend during Q2, our total cash liquidity, excluding customer deposits payable, remained at a high level of around RMB70 billion. With our robust cash reserves, we will continue to augment shareholder returns through active share buybacks to further enhance capital allocation and capital operation efficiency. As of the end of Q2, we repurchased around US$394 million worth of shares this year, which accounted for around 1.7% of the Company’s total shares outstanding at the end of 2024. We have consistently delivered on our promise to reward shareholders. Since the launch of our share repurchase program in September 2022, we have repurchased around US$2 billion in shares as of the end of June 2025, accounting for about 10.3% of our total shares outstanding before the program began. Today, we are pleased to announce that our board has approved an expansion of the existing share repurchase program. The authorization has been increased to US$5 billion, and the program has been extended to August 31, 2028. Going forward, we will continue to reward our shareholders who have grown with us and share the value we create. Despite fluctuations in the macro environment, we have delivered a top-line performance that significantly outperforms the market, underpinned by our solid business fundamentals and diversified portfolio. We are actively driving operational improvements to maximize the company’s long-term value. AI-driven refined operations and ecosystem optimization are continuously unleashing the platform’s long-term potential. Our healthy cash flow and proactive shareholder return policy demonstrate our firm commitment to long-term value creation. Looking ahead, we will join forces with all partners and shareholders to seize opportunities and create greater value together. Thank you. Next, I would like to turn the call over to our Chairman and CEO, Mr. Stanley Peng. Please go ahead, Stanley.

Thank you, Tao, for that overview of our updates for the first half of the year. Now I’d like to share the reasoning behind our initiatives and address some key interests and concerns. First, I’d like to talk about scale and efficiency. For our housing transaction services business, we have expanded our agent and store network very rapidly over the past few years. In the first half of the year, a large number of new brands, stores, and agents joined our network. However, we saw some softness in our efficiency indicators in housing transaction services in the second quarter. Our ACN and authentic listings were originally built to solve the key consumer pain points of that time. But as China’s real estate market evolved, consumer needs changed dramatically, and our responses didn’t fully meet those emerging needs. This created an urgent need for us to shift our growth engine from scale to efficiency. The main challenge was clear: how could we raise productivity per store and per agent, and increase platform efficiency, while maintaining the scale of our agent and store network? Resolving this will define the next stage of our development. To begin with, I want to stress that scale and efficiency are not a zero-sum tradeoff. Gaining efficiency doesn’t mean sacrificing scale. Reaching the current scale of our agent store network was no small feat. You’d be hard-pressed to find a comparable example anywhere in the world of a company that has reached such a large presence in a single city through its own operations. So how did we accomplish it? In the past, the scarcest resources in the market were high-quality assets and transaction security guarantees. That is why we built our ACN and introduced authentic listings and service commitments, integrating online platform innovation, offline business execution, and disciplined scientific management to support them. By providing what was scarce, we achieved a breakthrough in scale and built a deep competitive moat. So how do we move from scale to efficiency? First, by looking at our history, we need to find out the existing strengths that we can leverage while anchoring new capabilities for today’s context. Second, by responding to consumers’ evolving needs. In China’s changing real estate market, the new scarce resources are accurate market insight to help buyers make the right decisions, operational capabilities to help sellers market properties effectively, and the emotional value that comes from empathizing with customers. Buyers want professional advice, homeowners need skilled marketing support, and in the age of AI, customers crave emotional connection. Our task is to create clear pathways to deliver these values. As we seek a new growth paradigm, we start by challenging a paradox – large organizations often sacrifice efficiency to pursue scale. We already have the drivers for efficiency improvements. First, the transformation of customer demand acts as a natural form of selection. Second, AI-led innovation is delivering real productivity gains. As a new means of production, AI is becoming increasingly powerful, capable of replacing traditional means of production. These two forces enable us to raise efficiency while maintaining the scale of our agent and store network. Our endgame is clear, but how we get there is still being defined. Going forward, we will commit our energy and resources to these areas to reshape our growth path. Now I’d like to delve into the logic behind some of the business initiatives we are working on. I’ll start with the home renovation business. As Tao mentioned, our strategy centers on community-centric operations and our full-service premium store model. Our rationale reflects a major emerging trend. It is a shift from a traffic-driven mindset to a local community-centric approach. We have begun piloting this model through our first full-service home renovation premium store in Beijing. In the premium store, we put in showrooms with our modular renovation products so potential customers can see real, replicable home renovations based on typical local floor plans in the community. Our organizational structure has adapted accordingly. Designers, project managers, and workers are now dedicated to specific communities, gaining knowledge and experience of both properties and customers there. The key here is to “deepen our engagement,” locking in high-value areas and strengthening our presence to become customers’ first choice in the region, ultimately “occupying their mindshare.” Our goal is to bridge the distance between our services and users. We want to bring them closer physically, psychologically, and in decision-making, so we can evolve from a city-level renovation service provider to a community-level partner whose interests are deeply aligned with the customers’. Specifically: By opening our home renovation premium stores adjacent to our existing housing transaction contract signing centers, we’ve reduced the physical distance to our customers. This tells our customers we are a neighbor right down the street, not a large, distant company they have to drive an hour to reach. This builds trust and convenience at the same time. Our community-based premium stores and community-specific service providers are well-versed in the floor plans and customer needs, so they can offer tailored home renovation design plans, even before customers purchase homes. This shortens the decision-making distance for customers. Why is it? Because the biggest pain point in traditional home renovation is uncertainty. Customers often don’t know the final cost, what their homes will look like or whether the service will be reliable. We solve this with two innovative services: “community showroom designs” and “pre-signing measurement and drawing.” “Community showroom designs” say to potential customers: your neighbor’s home with the same floor plan as yours has already been renovated, with visible results, clear prices, and proven satisfaction. No guesswork is needed. Our “pre-signing measurement and drawing” flips the traditional “payment first, services later” model. We demonstrate value upfront by providing professional measuring and design renderings before customers commit. This gives people greater security and confidence right from the start. Our strategy also reduces the psychological distance for users. With a physical store, proven cases, and a professional consulting team in the community, our brand is no longer a cold, impersonal advertisement; it becomes a real, tangible “presence” customers can interact with anytime. The assets we gain from this kind of deep operation are more powerful than marketing, turning one-time “transactional customers” into long-term interactive “community users.” These are just some of the ways we think about our initiatives in the home renovation business. Next, I’d like to talk about how we view the home rental business. Tao has already covered much of our progress in this segment, including product iterations and the broader use of AI to boost property managers’ efficiency and streamline other operations. Why do we pursue maximum efficiency in this business? Because under the traditional management model, our Carefree Rent business inevitably faces diseconomies of scale. Once the number of units reaches a certain level, complexity may rise sharply, due to the non-standard nature of our products, the service provider abilities, and sales negotiations. At the same time, the rental business operates on thin service fee profit margins, which cannot absorb losses from non-standard operations and low efficiency. These three major challenges form an “iron triangle” that compels us to break through the traditional model and pursue maximum operating efficiency. So how do we achieve maximum efficiency? In Phase One, we restructured our organization, moving away from the “all-in-one manager” model to six specialized roles. Fully aligned with the logic of our ACN, this division improved professional skills, reduced service variance, and embedded these capabilities into our platform. In Phase Two, we optimized our product model, shifting from the high-risk, non-standard, vacancy-prone “lease-out” model to a steady rent pass-through model with unified service fees. This stabilized revenue per property, aligned team goals, and removed the obstacles for scaling growth. We also digitalized processes through our SaaS system, accumulating structured data to fuel AI applications. In Phase Three, we began to build intelligence into operations, deploying our AI-human model, where AI handles standardization, pricing, auditing, and 24/7 virtual service. We hope that AI could cover 80% of standardized work, while people can focus on trust, irregular cases, and high-quality service. The logic tying all of this together is about transforming a non-standardized offline industry full of uncertainty into a data- and intelligence-driven business with more certainties. At its core, our system reduces reliance on individual experience, smoothing out fluctuations in service quality and customer experience. Efficient operations and consistent service quality create a growth flywheel, reinforcing the synergies across the home rental business, home renovation, and housing transactions. Our vision is to build an AI-driven rental platform by combining AI, IoT hardware, and operating processes. We hope this platform will give the rental industry a proven and scalable profit model. Meanwhile, the pursuit of operational excellence will inevitably compel our whole organization to develop more efficient operational mechanisms. We also hope it can be an example for the traditional service industry, showing how structure, model design, and technology can solve issues like “non-standardized nature” and “diseconomies of scale.” Now moving to our Beihaojia business. Why are we pursuing this business? We will not be developers; to be clear, we will not be adopting an asset-heavy model. The traditional real estate development business used to depend on land and money. In today’s market, there is a new variable: a customer-oriented mindset. Because we are so close to customers, we can collect more customer insights and data to add value to this third variable. In the early stage of this business, we ran two self-operated projects to test our understanding and to see what value we could create for this third factor. Genuine customer needs are at the very core of our product design and construction. We leveraged our robust data and AI-powered capabilities, including pricing prediction, unit mix optimization, and potential customer insights, to deeply understand our target customer needs. Our project positioning, product design, and construction adhere to these authentic customer needs, including many small details traditional developers might overlook, but that we consider critical to the long-term living experience. In our Chengdu Beichen Project, we have dedicated meticulous design and construction efforts to over 108 quality-driven details that may seem minor yet meaningful. These span from urban integration of architectural design, landscape planning, homecoming journey experiences, interior spatial planning, to AIoT-enabled sensory systems covering sight/sound/smell/taste/touch/consciousness, as well as lifestyle scenarios and property services. This productization capability is becoming crucial as the market shifts to buyers. This means the supply side must offer differentiated, not homogeneous products. And this is how we add value to the industry through the third variable of production, beyond land and money. Finally, we now stand at a crucial turning point. Balancing scale and efficiency, adapting to evolving customer demands, and keeping pace with rapid technology development are all issues we must address. We have already started exploring and testing new approaches across our businesses. While maintaining the scale advantages of our platform, we aim to revamp our service interface through community-centric operations, unlock organizational efficiencies with AI, rebuild our product logic with a customer-centric mindset, and continuously shape new paradigms in the residential services industry. This concludes my prepared remarks for today. Operator, we are now ready to take questions.

Thank you. As a reminder, we only accept questions in English. Please limit yourself to one question, and if you have more, you can re-enter the queue. If you ask a question in Chinese, please provide an English translation. Your first question comes from Timothy Zhao from Goldman Sachs.

Thanks management for taking my questions and congrats on the solid results and very excited to hear about your new approach in growing the business in the future. My question is on the secondary home. I just wondering if management can provide us any overview on the Q2 secondary home market and how should we expect the trajectory into the second half of this year? What kind of policy tools that we can expect for the rest of this year now? Thank you.

Thanks for your question. Let’s first take a brief look at the market in H1. The total value of housing transactions nationwide was stable overall. The market was off to a good start in Q1, sustaining the recovery momentum from Q4 2024, but both the number and price of transactions weakened significantly in Q2. Divergences also intensified. Based on NBS data, new home sales nationwide dropped by 5.2% year-over-year in H1. CRIC research indicates that the TOP 100 developers saw steeper declines, posting a 10.9% H1 sales slump that intensified to 14.5% year-over-year in Q2. The existing home market held up relatively well. According to Beike Research Institute, in H1, the total value of online registered transactions for existing homes rose 8.3% year-over-year. This was driven by a 19% increase in the number of transactions, even though average prices fell by 9%, both year-over-year. That said, momentum slowed in Q2. The transaction value growth rate dipped to just 2% and the number of transactions also slipped to a 12% gain. In June, the number of transactions decreased by 8% year over year, and home prices dropped further month-over-month by 1.7%. Rent has been more stable than home prices. Nationwide rental yield has been steadily rebounding since 2021. In June this year, it reached a high of 2.5%, about 40% higher than its lowest point, creating value support for home prices. Structurally, existing homes continued to perform better than new homes. Existing home transactions tended to have a larger average GFA, lower total price, and a higher share of nearly-new existing homes. Population movement from lower- to higher-tier cities remained steady, and we continued to see wide differences across cities in terms of housing price, inventory sell-through cycles and land auction premiums. Our key indicators, like the homeowner price adjustment prosperity index and the agent confidence index, showed weak market sentiment. The fading effects of the new policies, China-US trade tensions, short-term policy vacuums, and seasonal market corrections sapped market momentum. Reinforced expectations for price declines constrained market recovery. All of this added to strong downward pressure for the market. Since the beginning of July, the market downturn has picked up speed. The number of existing home transactions fell by over 5% month over month, while prices dropped 1.5%. New home subscriptions fell 25% month-over-month, signaling a period of sharp correction. Looking ahead, the market's path will still depend on the pace of future policies and the supply-demand balance improvement. Both are key to restoring confidence, which in turn influences buying behavior and price trends. On June 13, the State Council Executive Meeting once again emphasized the need to “double down on stabilizing the market,” signaling the potential for stronger policy support. On top of the current policies that remain in place, there is still room for new, stronger policies aimed at boosting demand and improving supply. On the demand side, leading cities have further relaxed purchase restrictions. Urban renewal, relocation vouchers, and purchase subsidies can also unlock potential home-buying demand. On the supply side, high-quality housing from fourth-generation houses, and nearly new existing home supply enabled by the easing of sales restrictions, could elevate supply-side quality, and potentially market sentiment and transaction volume. Proactive policies can help counteract the market's downward trend and support a shift toward recovery. Thank you.

Your next question comes from John Lam from UBS.

Thank you for taking my question. Could management provide insights on what steps have been taken in light of the downturn in the property sector to appeal to investors? For instance, regarding market share, agent productivity, or store performance. Additionally, how does management view the growth strategy for both agents and the number of stores? Thank you.

Thanks for your question. We hope to outperform the market for a long time to come, and have taken steps to achieve that. I’ll answer your question from two angles: scale and efficiency. First, the continuous expansion of our platform’s agent and store network over the past few years has fueled fast business growth. Going forward, we will slow the pace of our store and agent growth and focus more on efficiency for sustainable development. This strategy will vary by city. In places where store network coverage is already high, we will impose higher quality and ROI requirements to onboard new stores. In a few cities where store network coverage is still relatively low, we’ll continue to make strategic investments. By the end of the year, we expect to keep our store and agent numbers stable, outside of Beijing and Shanghai. In those two cities, where agent growth has been strong over the past few years, we are consolidating lower-performing stores and phasing out lower-performing agents. I believe it’s time to shift our growth focus from scale to efficiency, starting with deeper operational efforts in the short term. In the long term, we expect technology to drive industry-wide gains in total-factor productivity. In details, we will implement in-depth, systematic scientific management operations, particularly by enhancing operations that add competitive dimensions. This involves managing processes that influence competitive outcomes, such as focusing on properties, customers, and improving collaboration and matching, to enhance the competitiveness of individual stores on the platform. We will also continue to strengthen our key operations initiatives, such as our points-based store incentive system, regional co-governance councils, management of high-quality business districts, and the separation of agents’ roles for homeowners and buyers, which we piloted in Shanghai. These projects are helping us improve our ecosystem and guide service providers’ behavior. For example, stores in high-quality business districts had 1.44 times the average productivity of other stores, compared with less than 1.4 times in the second half of 2024. Over the long term, technology will be key to enhancing our efficiency and consistently delivering alpha. We see that advances in large AI model technology, combined with our unique scenarios and data in the residential services sector, have strong potential in reshaping user experiences, boosting efficiency, and spurring transformation in the real estate industry. We have established an AI project matrix that develops different AI applications simultaneously for different roles on the C-end and B-end, on both the strategic and operational levels. Some of these applications have demonstrated good results. One example is our AIGC marketing and AI-driven CRM products that address customer acquisition and conversion for agents. We have built an intelligent AIGC marketing agent for real estate agents to support customer acquisition through self-media, private domain traffic operations, and lead conversion. It offers a full set of tools, helping agents create multi-modal content for multi-channel customer acquisition, perform better data analysis and lead identification, analyze leads and create ice-breaker scripts, and generate price trends automatically. Our intelligent AI-powered CRM agents strengthen customer acquisition and conversion for all customer-facing roles. Here, I'll refer to our brokerage agents as "brokers" to avoid confusion. For existing home brokers, our AI CRM improves customer management through a constantly evolving multi-agent system, powered by our user data. It gives brokers personalized guidance ranging from market insights and customer strategies to recommended actions. This input helps brokers understand users’ needs, gauge their intentions, and spot new opportunities. It also automates personalized follow-up tasks to drive transactions. By the end of June 2025, “Like”, one of our AI-driven CRM product applications, was in use across 69 cities, empowering over 335,000 brokers. The product’s penetration rate exceeded 75% across Beijing and Shanghai. In Xi’an, for example, brokers who use Like extensively achieve a 30% higher conversion rate for formal client mandates, and about a 20% higher conversion rate for showings, compared with brokers that use the product less frequently. For new homes, we have AI-driven CRM agents “Qianji” and “Qianzhi” that help new home sales managers improve listing management efficiency and strengthen both matching and marketing for new home products. On the consumer side, for example, our Pudding AI online service assistant now provides real estate market analysis, city-wide house searches, regional analysis, home listing comparisons, and preliminary matching of home listings and agents. We began gray-box testing in May, making Pudding available to select users in 11 cities. In July, its MAUs reached 780 thousand, up 10% from June. Conversation volume grew by 59% and average time spent per user was up 14%. Pudding now supports multi-language, multi-modal services. We are also transforming Pudding to allow it to proactively explore customer needs and handle tasks, while providing more precise, high-quality, instructive answers. These enhancements will help users make decisions and move transactions forward. Pudding is our first to-C trial in offering AI-powered services, and we are making extensive efforts to offer more intelligent services to the industry. We believe AI is the most crucial driver for our next-generation productivity improvements. We will keep you posted on our internal progress.

Your next question comes from Griffin Chan from Citi.

Thank you to management for addressing my question. I would like to know how the new property development model, including property sales or the promotion of quality housing, will create new opportunities for Beike, particularly in areas like demand forecasting or product design. Thank you.

Thank you, Griffin. This year, supply-side policies in real estate have accelerated the push for better living quality, especially through “high-quality homes” and move-in-ready new homes. These measures are being implemented at a faster pace. Residential products that meet new national standards have performed well, and the pilot sales of move-in-ready new homes in Xinyang have set a good example. Surveys show that among the factors holding back buyers, price expectations account for nearly 50%, while home suitability accounts for about 20%. As new home products better meet home upgraders’ suitability demands, and the system for selling move-in-ready new homes gradually expands, we expect to see reduced quantity and improved quality supply. These new models set much higher requirements on developers, from securing funding and ensuring project returns to understanding upgrade needs, project positioning, pricing, and sell-through marketing. This will further highlight the value Beike brings to developers: In terms of impact on the current brokerage model: In the short term, in first-tier cities, new home products that meet the new standards will have a lower brokerage sell-through proportion and commission rate than products under the old standard. Nevertheless, in most cities, new-standard products currently only account for around 10% of units, and their presence will push old-standard products to raise their brokerage services penetration ratio. As a result, the overall impact on the brokerage channel sales market is small. When new-standard products make up over 30% of the project launches in a city, for example as they do in Xi’an, their brokerage penetration and commission rates will match those of old-standard products. The fast sell-through of new-reg products can boost agents’ confidence in new home products, forming a virtuous cycle. Opportunity for cooperation model upgrade: The industry’s new models will also drive upgrades in how we work with developers beyond the brokerage channel sales model. Our goal is to offer tailored services for managing the full project lifecycle, based on each developer's needs and project type. This will further highlight the value of our Beihaojia business, from its C2M product solutions to its integrated online-offline marketing services. Price forecasting capability: We leverage systematic modeling approaches, including a subjective factor-eliminated pricing model with rolling review and calibration mechanisms, and comparable price trend analysis based on authentic existing home market data. Our algorithms strip out structural factors, so that prices are comparable both across markets and over time. This pricing capability helps developers objectively assess price trends and set accurate prices, avoiding profit loss from mispricing. When competition in selling move-in-ready new homes intensifies, it can also improve a project's value-for-money positioning. For example, in Nanjing and Wuhan, new-standard products are noticeably more competitive, exhibiting independent pricing trends. Beike's price forecasting capabilities enable granular segmentation analysis to better characterize such market dynamics. Today, our pricing model already has a fairly high level of accuracy. Regarding the unit mix forecasting capability, we apply machine learning algorithms to model and forecast customers’ housing unit needs, drawing on both potential customer behaviors and historical transaction data. In June 2025, the compliance rate of our sample simulations continued to rise. Our goal for 2026 is to have full plot and market coverage to help developers plan unit mixes more accurately, avoid inventory buildup, and speed up sales of move-in-ready new homes. Regarding the customer insights, we can clearly define and pinpoint potential customers for building projects in specific districts, along with their needs and profiles. This includes identifying their purchasing power, preferred housing unit type, location, age, purchase purpose, family size, and so on. Using our potential customer model, we can forecast high-purchase-intent buyers in the next 90 days for any given district and their specific needs. This gives developers the information they need to enhance competitiveness by targeting the right demographics and designing or optimizing products that meet the new standard requirements. At the current stage, we hope to help developers position their products accurately at the early stage, reducing the cost of later-stage adjustments and improving product alignment with market demand. Looking ahead, we will focus on building customization and community operation capabilities to help developers stand out in the move-in-ready new home market. By complementing developers with our strengths in move-in-ready new home sell-through, regulatory adaptation, and bottom-line protection, our value to developers will extend from brokerage to product source.

Your next question comes from Daniel Chen from JPMorgan.

Thank you to the management for taking my question. My question is regarding the Home Renovation and Furnishing business. We have observed strong margin improvement year-over-year and healthy revenue growth. What are the main drivers behind this growth? Is there still potential for cost optimization? Additionally, will we expand our city coverage or focus on optimizing our store network? Thank you.

Thank you, Daniel. The home renovation and furnishing business maintained a relatively high growth rate in the first half of the year. Scale-wise, its revenue reached RMB7.51 billion, up 16.5% year-over-year. On the profitability front, the segment’s profit margin was 32.3% in the first half of the year, rising by 1.3 percentage points from the same period last year. Operational efficiency also improved significantly at the city level in Q2. “Operational efficiency enhancement” is the focus of our home renovation business this year. We have implemented a series of initiatives aimed at enhancing fundamentals like our product and delivery capabilities, streamlining our organizational structure, and amplifying management and operational efficiency. These efforts have led to continuous improvements in our performance. Here I’d like to elaborate. In terms of costs, the home renovation business’s expenses mainly include material and labor costs. For materials, we cut procurement costs by consolidating brands and SKUs and moving to centralized procurement. By leveraging customer insights to streamline brand selection and SKU counts, we have consolidated procurement to three or fewer brands for most categories, and achieved significant SKU reduction. Centralized procurement rate for primary and auxiliary materials reached more than 60% in Q2 2025, compared with over 20% in the same period last year. The increased procurement volume per SKU in each category has led to a significant decrease in the unit price of some products that were awarded contracts. On the labor side, we have improved service providers’ work efficiency by optimizing our order-dispatching rules, enhancing the system's order-distribution capabilities, and focusing project managers’ service areas on specific business districts. Also, we are allocating more resources to high-quality service providers. In Q2 2025, average monthly order intake per professional project manager was more than double last year's average, driving significant improvements in both their efficiency and income. In terms of sales and marketing expenses, these mainly cover sales personnel costs for designers and customer managers, and store and brokerage channel costs. As a percentage of revenue, sales personnel costs for designers and other roles significantly decreased year-over-year, primarily due to our agile transformation of the home renovation business’s organizational structure. We streamlined the designer team, optimized the overall home renovation business workflow, eliminated certain roles in the sales process, and shifted the functions toward agents. Our digital tools, such as AI Proposal and Lightweight BIM, have significantly improved operational efficiency in the sales process. The average monthly order volume per designer has increased from around 0.8 orders last year to over 1.2 orders in the second quarter of this year. To optimize store costs, we closed some underperforming large stores and piloted small-sized premium home renovation stores near our home transaction contract signing centers, where we integrated master-designed productized showrooms. We hope to strengthen the synergies between our housing transaction business and new initiatives, reducing store costs and improving sales per unit area, while exploring a new one-stop full-service home renovation model. For G&A expenses, we also made some structural enhancements, expanding the service scope of middle- and back-office personnel. The average number of orders supported by each middle- and back-office personnel increased by 70% year-over-year in Q2. These initiatives enabled us to achieve better operational results while continuously improving service quality. In Q2 this year, the customer complaint rate of our home renovation and furnishing business dropped to below 10% from over 25% in the same period last year. We have seen a significant improvement in Unit Economics (UE) at the city level. In the Beijing area, revenue increased from over 700 million yuan in the first half of 2023 to 1.15 billion yuan in the first half of 2024, and exceeded 1.5 billion yuan in the first half of this year, representing year-on-year growth of over 30%. The gross profit margins in the first half of 2023, 2024, and 2025 were 35%, 36.1%, and 36.3% respectively. The operating profit margin at the city level rose from around 5% in the first half of 2023 to over 11% in the same period of 2025. The remarkable improvement in model cities’ UE has boosted our confidence in our business’s continuous operation and future success. There is still plenty of room for operational efficiency improvements. Moving forward, our focus will shift from organizational structure optimization this year to implementing ongoing innovation in business models, products, and technology.

Your next question comes from Xiaodan Zhang from CICC.

Thank you, management, for addressing my questions. Since its launch at the end of 2023, Beihaojia has introduced several projects to the market. Based on the operational experience gained in the past year, could you provide insight into the future plans for Beihaojia? Specifically, what business model will it follow, and will there be a cap on the investment budget for individual projects? Thank you.

Thank you. We’ve been very clear about Beihaojia’s strategic direction. We are adamant about not being developers. In terms of business models, we are dedicated to an asset-light business model. Except for our Chengdu Financial City project and Shanghai Fengxian Xincheng project, we will not independently operate other projects. The C2M models do not provide funding solutions. We will continue to explore a platform model that offers full front-to-back-end services for developers, construction contractors, property owners, and other partners, including product solutions and marketing services. Our product solutions cover C2M product positioning and design plans backed by AI and big data. Our marketing services are integrated online-offline promotional services for more efficient customer acquisition. Regarding tied-up capital, which I know concerns all of you, we have set strict limits on the peak total investment from our own funds. Based on the amounts already deployed by the Group as of June 30, we will invest no more than RMB 1 billion in additional self-owned funds. After the exit of the two proprietary development projects in Chengdu and Shanghai, the limit of our capital occupation for this business will be reduced by the investment amounts of these two projects, further lowering our aggregate self-owned funding cap. Thank you.

We are now approaching the end of the conference call. I will now turn the call over to your speaker host today, Ms. Siting Li, for closing remarks.

Thank you once again for joining us today. If you have further questions, please feel free to contact Beike’s investor relations team through the contact information provided on our website. This concludes today’s call, and we look forward to speaking with you again next quarter. Thank you, and goodbye.