KE Holdings Inc. (贝壳 Beike): Can China's Housing Super-App Out-Build the Property Crash?
I. Introduction & Episode Roadmap
On August 21, 2026, KE Holdings Inc.—widely known across China as 贝壳 Beike, listed as 2423.HK in Hong Kong and BEKE on the New York Stock Exchange—reported second-quarter results that appeared internally contradictory. Gross transaction value across its platform rose 6.3% year over year to RMB 933.8 billion, even as revenue fell 5.7%. Yet net income doubled, rising 100.8% to RMB 2.62 billion.12
More volume, lower revenue, and twice the profit.
The company behind those numbers is China's dominant housing transaction platform: dual-primary listed in Hong Kong and New York, headquartered in Beijing, and accessible online at bj.ke.com and investors.ke.com.25
That operational arithmetic reflects the modern KE Holdings story. Over the past five years, the company has navigated a series of severe external disruptions—the death of its founder, an antitrust inquiry, a short-seller report, and a multi-year downturn in Chinese residential property—responding by restructuring its internal cost dynamics. Whether this operational pivot establishes a durable competitive moat or merely represents a well-managed retreat remains the central question facing investors.
The company's public history began under very different market conditions. KE Holdings listed on the NYSE in August 2020, raising $2.12 billion at $20 per American depositary share—priced above its offer range in the largest U.S. listing by a Chinese issuer since iQIYI—with backing from 腾讯 Tencent, SoftBank, 高瓴资本 Hillhouse Capital, and Sequoia Capital China.3 Nine months later, founder 左晖 Zuo Hui died at age 50.4 Days afterward, Reuters reported that Chinese market regulators had opened an antitrust investigation into the company; Beike called the report "fake news" in a single-sentence WeChat post.7 Seven months after that, a short seller published a report alleging the company had fabricated its core operating metrics.8 Compounding these corporate events, China's residential property market entered a deep structural contraction that has persisted for five years.
Despite these headwinds, KE Holdings remains profitable and financially resilient. The company repurchased approximately $460 million of its own stock in the first half of 2026 alone, bringing total buybacks to about $2.99 billion since the program began in September 2022—equivalent to roughly 14.8% of the shares outstanding before it started.13 It maintains a dividend, backed by a balance sheet holding approximately RMB 67.3 billion of broad cash, excluding customer deposits.13
This raises a key strategic question: is Beike a genuinely moated platform—a data and network asset capable of compounding market share as a shrinking industry consolidates—or is it a well-executed brokerage whose peak growth ended at listing, now taking on property-development exposure that threatens the asset-light balance sheet that defined its original model?
Understanding Beike's trajectory requires examining 链家 Lianjia's origins in Beijing's fragmented 2001 brokerage market; the 2018 decision to share its proprietary agent network and transaction data with competitors; its August 2020 IPO near the market peak; the passing of 左晖 Zuo Hui; the severe operational shocks of 2021–2022; the core transaction engine that continues to generate the majority of operating profits; its expansion into home renovation and rental management alongside newer, higher-risk ventures; CEO 彭永东 Peng Yongdong's capital allocation track record; and the underlying conditions required for either the bull or bear thesis to hold.
The journey began with an insurance agent in Beijing who sought to reform a real estate brokerage market long defined by opaque practices.
Subscribe on LinkedInII. Origins: Zuo Hui and the Lianjia Bet (2001–2017)
Beijing's residential brokerage market in 2001 functioned less like an organized industry than an information arbitrage. Listings were frequently fake—posted below market rates to generate leads, then withdrawn once a buyer engaged. The same apartment often appeared across half a dozen agencies at different prices. Agents frequently represented both sides of a transaction without disclosing full terms to either. Crucially, no authoritative public record existed to verify which buildings stood, how many units they contained, what properties had traded for, or whether a seller genuinely held title.
Into that environment entered Zuo Hui. Born in Shaanxi province, Zuo graduated in 1992 with a chemical engineering degree from Beijing University of Chemical Technology, earned his initial capital in sales and insurance, and founded Beijing Lianjia Real Estate Brokerage in 2001.6 KE Holdings dates its operating history to that launch, citing "over 24 years of operating experience through Lianjia since our inception in 2001" in its FY2025 annual report.2
Zuo's core insight was less technological than diagnostic: the market's breakdown stemmed not from technical inefficiency, but from misaligned incentives. Dishonesty was the profit-maximizing strategy for individual participants. Breaking that equilibrium required building a brokerage where misrepresentation was structurally impractical. That meant employing agents directly rather than franchising, enforcing fixed and transparent commission splits, and constructing an independent database of physical properties that agents could not falsify.
That database became the 楼盘字典 Housing Dictionary. Beginning in 2008, Lianjia deployed teams door-to-door and building-by-building to compile a private census of China's urban housing stock. Beyond addresses, the survey recorded unit layouts, window orientations, ownership statuses, and transaction histories. It was a capital-intensive, labor-heavy endeavor that initially appeared to be an unremunerative use of brokerage capital.
In May 2011, Lianjia leveraged the database to launch a "genuine listings" standard—guaranteeing that advertised properties actually existed, were available for sale at the listed price, and featured authentic photographs—backed by financial compensation for customers who identified fake listings. Competitors could replicate the marketing slogan, but without a verified database, they could not underwrite the guarantee.
The high cost of Zuo's direct-operation model was central to its strategic logic. Most Chinese brokerages favored franchising: licensing a brand, collecting fees, avoiding payroll, and accepting limited operational control. Direct operation inverted those economics. Lianjia carried agents directly on its balance sheet, paid fixed commission splits, and absorbed the costs of data collection and training. In a booming property market, franchisors generated higher returns on capital and scaled faster. Zuo accepted lower capital efficiency for over a decade in exchange for strict operational control and customer trust—an asset that would only realize its full value when scaled beyond Lianjia's proprietary storefronts.
By the end of 2025, the Housing Dictionary covered approximately 303 million properties.2 That dataset represents a formidable moat. While competitors can hire agents, lease storefronts, and build mobile applications relatively quickly, retroactively conducting a multi-year physical survey of Chinese housing stock and aggregating its transaction history is far more difficult to replicate.
Everything Lianjia constructed between 2001 and 2017—direct management, standardized commission splits, the housing census, and verified listings—served as infrastructure to resolve systemic market mistrust. Yet by 2017, Lianjia remained a traditional brokerage whose growth was constrained by its physical footprint and agent headcount, operating alongside hundreds of thousands of independent agencies it did not own.
That bottleneck posed a fundamental strategic question: if the company's core asset was its trust-enforcing infrastructure rather than individual brokerage transactions, why restrict that infrastructure to its own internal deals?
III. The Platform Pivot: Beike and the Agent Cooperation Network (2018)
In 2018, the company made a strategic pivot that appeared counterintuitive at the time: opening its Housing Dictionary, listing standards, transaction workflows, and technology stack to rival brokerages—including competitors Lianjia had spent 17 years fighting—and charging them for access.
Beike launched that year as an open platform.2 The mechanism driving this transition was the Agent Cooperation Network (ACN), which served as the company's central operational engine.
Before ACN, residential transactions in China rewarded only the closing agent. That structure incentivized agents to hoard listings, as sharing meant surrendering the full commission. Listing hoarding produced duplicated entries, misleading prices, and widespread consumer distrust, which in turn depressed transaction velocity across the market—creating a coordination failure locked in a persistent bad equilibrium.
ACN unbundled the transaction into distinct operational roles: listing acquisition, property verification and photography, owner relationship management, buyer matching, property showings, and deal closing. The platform established standardized commission splits for each role in advance. Once commission sharing became a standardized platform rule rather than an individual negotiation, sharing listings became rational, ensuring agents were compensated for sourcing properties even if another agent closed the deal.
This structure operated like a public highway system with a standardized metering framework, replacing a fragmented network of competing private toll roads. Under ACN, the Housing Dictionary served as the underlying verification system, ensuring commissions were not split on non-existent or disputed listings.
The platform operated across three distinct brand tiers. Lianjia served as the company's directly operated brand, functioning as the operational quality anchor. Relaunched in 2018, 德佑 Deyou operated as the franchise brand for smaller independent brokerages seeking brand recognition and standardized rules without full integration. Beyond these, connected brands represented independent third-party brokerages that plugged into ACN and paid platform service fees, with the company collecting additional franchise fees from stores operating under its franchise banners.2
Operational data illustrates the scale of network adoption. As of December 31, 2025, the platform carried approximately 523,000 agents across roughly 61,000 stores, representing 279 real estate brokerage brands.2 Lianjia accounted for just over 90,000 active agents across approximately 4,900 active stores.2 Consequently, the company directly employed roughly 17% of the agents and operated less than 8% of the storefronts on its platform. In its FY2025 annual report, management noted that connected brands, combined with specialized sales channels for new home transactions, "contributed a substantial majority of the GTV on our platform in 2025."2
In strategic terms, this model combined two distinct competitive advantages. First, it generated network economies: each additional participating brokerage deepened shared property inventory, increasing platform utility for all participants. Second, it leveraged the Housing Dictionary as a cornered resource—a proprietary data asset compiled through years of physical field surveys that competitors could not easily replicate.
However, ACN served as a cooperation protocol rather than a price-setting mechanism. While it dictated how commissions were divided among agents, it did not dictate total commission rates—a distinction that became critical as Chinese regulators subsequently turned their attention to overall transaction costs. Additionally, relying on external partner networks introduced operational risks. In its risk disclosures, the company acknowledged that with a growing number of participating brands, it "may be difficult for us to effectively monitor and control these brands and agents," warning that violations such as bypassing the platform to avoid commission splits could "diminish" ACN's effectiveness.2
This structure differentiates Beike from U.S. proptech models. Companies such as Zillow and Compass primarily monetize lead generation and advertising, operating upstream of the transaction. In contrast, Beike manages the end-to-end transaction lifecycle, controlling listing verification, commission splits, contract workflows, escrow services, and closing support. This transaction-level integration generates higher take rates and deeper agent retention, positioning the platform as an essential operating system. However, it also requires a heavier cost structure and expands the company's regulatory footprint.
By 2019, the platform's market adoption was highlighted when Century 21 China, operating through IFM Investments, joined the network—remaining a related-party platform-services customer through FY2025.2 A major competitor's franchise network had concluded that integrating into Beike's infrastructure offered greater value than competing against it.
Two years later, the company presented that platform model to public investors.
IV. Going Public Into a Storm: The 2020 IPO and Capital Base
In August 2020, KE Holdings completed its initial public offering on the New York Stock Exchange, raising $2.12 billion at $20 per American depositary share—above its indicated price range—in what was then the largest U.S. listing by a Chinese issuer since iQIYI.3 The stock surged on its trading debut.21
Within months of listing, public markets capitalized the platform above China's largest property developers, valuing a brokerage and data services provider higher than the homebuilders whose inventory it sold.20 Strategically, that valuation disparity reflected fundamental operational differences: Beike operated without land reserves, construction obligations, pre-sale liabilities, or debt-heavy balance sheets. Investors were pricing in a structural transition, wagering that China's real estate market would evolve from a construction boom into a transaction-driven service economy where an asset-light intermediary commanded higher valuation multiples than physical developers.
However, that thesis assumed a shift in market composition rather than an overall industry contraction. Instead of a structural rotation from property development to brokerage transactions, the broader real estate sector subsequently experienced a synchronized downturn across both segments.
Beyond market dynamics, two governance features defined the company's capital structure. The first was a dual-class voting system under which Class A common shares carry one vote per share, while Class B shares carry ten votes. As of February 28, 2026, holders of 138.6 million Class B shares controlled 29.9% of aggregate voting power, concentrating effective governance authority in the hands of co-founders 彭永东 Peng Yongdong and Shan Yigang.2 Consequently, while public shareholders held the majority of economic equity, corporate control remained firmly with insider leadership.
The second structural feature was the variable interest entity (VIE) framework. Due to Chinese regulatory restrictions on foreign ownership in value-added telecommunications services, the offshore Cayman Islands holding company does not directly own the licensed domestic operating entities, controlling them instead through contractual agreements. In its FY2021 annual report, the company disclosed five primary VIEs—Beijing Lianjia, Tianjin Xiaowu, Yiju Taihe, Beijing Beijia, and Beijing Beihao—which collectively held 29.5% of group cash and 12.9% of total assets at the end of that year.14 While standard for foreign-listed Chinese technology firms, the arrangement carries ongoing legal risk; management acknowledged in corporate filings that an adverse ruling by Chinese authorities could force the deconsolidation of these operating entities.2
In November 2020, three months after its initial listing, KE Holdings completed a follow-on equity offering. Disclosures in those prospectus documents regarding total platform store and agent counts as of September 30, 2020, subsequently became central to legal challenges—eventually forming the sole surviving claim in shareholder litigation following a short-seller report.2
Having raised public capital, established its insider control structure, and secured a valuation premised on a maturing transaction market, KE Holdings entered 2021 with substantial financial liquidity. However, impending macroeconomic headwinds and regulatory shifts were about to test the stability of that foundation.
V. Losing the Founder: Zuo Hui's Death and the Succession Test
On May 20, 2021, KE Holdings announced the death of founder Zuo Hui at age 50, attributing it to an unexpected worsening of an illness. At the time of his passing, he was reportedly China's fifteenth-wealthiest person, with an estimated net worth near $14.8 billion.46
Zuo's sudden departure created a profound operational and strategic void. He was not a figurehead chairman, but the author of the company's operating philosophy—the leader who determined that a brokerage should conduct a physical housing census, that fake listings deserved zero tolerance, and that a dominant proprietary dataset should be shared with direct competitors. A corporate culture built around one founder's contrarian judgment faces immediate exposure once that leadership disappears.
The board acted swiftly to lock in management stability. Within days, co-founder and chief executive 彭永东 Peng Yongdong was appointed chairman, chief operating officer 徐万刚 Xu Wangang joined the board, and Zuo was designated chairman emeritus.5 Peng emphasized operational continuity, reassuring investors that the company would "move forward to execute on our strategy and growth initiatives."6
That leadership transition serves as a critical lens for assessing the company's subsequent trajectory. Although Zuo established the platform's foundation, every defining event of the modern investment case—the regulatory inquiry, the short-seller defense, the property downturn, the 圣都 Shengdu acquisition, the Hong Kong dual-primary listing, the expansion into home renovation and rental management, the launch of the 贝好家 Beihaojia development unit, the stock repurchases, the dividend program, and the 2026 efficiency pivot—has unfolded under Peng. Evaluating KE Holdings based solely on its founder's legacy overlooks the realities of the current executive team.
The board's immediate actions reflected a deliberate commitment to the existing strategy. The directors conducted no external search for a new chief executive, launched no formal strategic review, and placed no pause on platform initiatives. Instead, the chairmanship passed to the sitting chief executive while the broader operating leadership remained intact, with Zuo retained symbolically as chairman emeritus.5 This rapid consolidation of authority prioritized stability when market anxiety ran high, though it also bypassed an opportunity to re-examine strategic assumptions before major industry headwinds materialized. The central question for investors was whether a management team accustomed to executing Zuo's vision could adapt independently when market conditions deteriorated.
The first major test of new leadership arrived two months later. In July 2021, Peng announced an agreement to acquire Shengdu, a Hangzhou-based home renovation business, capping initial transaction consideration at RMB 8 billion. By the time the transaction closed in April 2022 against a backdrop of slowing housing sales, management had renegotiated the purchase terms down to RMB 3.92 billion in cash plus 44,315,854 Class A ordinary shares.215
The transaction reflected both pragmatic capital discipline and shifting operational realities. While the revised price demonstrated management's willingness to adjust terms as macroeconomic conditions deteriorated rather than overpay for asset-light expansion, it also underscored the weakening fundamentals of target assets in a contracting property market. That willingness to renegotiate terms provided an early signal of how Peng's team would approach capital allocation as industry stress intensified.
Yet any momentum from that operational adjustment was quickly eclipsed. Within eighteen months of Zuo's passing, leadership confronted three simultaneous crises—each capable of reshaping the company's future on its own.
VI. The Triple Shock: Antitrust, Short-Seller Attack, and the Property Crash
Shock one: the regulator. On May 25, 2021—five days after Zuo Hui's death—Reuters reported that China's State Administration for Market Regulation (SAMR) had launched a formal inquiry into whether KE Holdings forced real estate developers to list exclusively on its platforms, a practice known in Chinese antitrust parlance as "choose one from two." The enforcement tactic had produced a record RMB 18.2 billion fine against Alibaba one month prior. Beike responded via a single-sentence WeChat post calling the report ungrounded, without providing further detail.7
The scrutiny arrived alongside the broader 2021 tech platform crackdown and the common prosperity (共同富裕) policy initiative, which repositioned major consumer platforms as instruments of public policy rather than autonomous commercial actors. For an intermediary operating between developers, agents, and buyers on the single largest household expense in China, regulatory scrutiny posed a fundamental business risk.
The ultimate resolution of the inquiry differed markedly from initial public speculation. No public enforcement order, fine, or official finding against KE Holdings regarding exclusive dealing was ever disclosed. In its FY2025 annual report, the company's legal proceedings section cited only shareholder litigation, while framing antitrust compliance entirely as a prospective risk—noting that non-compliance, if formally established, "may" subject the platform to administrative penalties.2 While the company avoided a direct fine, regulatory pressure subsequently re-emerged through structural policy guidance.
In April 2023, the Ministry of Housing and Urban-Rural Development and SAMR jointly issued guidelines for real estate brokerage services, urging agencies to lower transaction fees and encourage equal fee-sharing between buyers and sellers. The company detailed its compliance in public filings: "We have made corresponding adjustments in certain cities. For example, in September 2023, we reduced commission rate for existing home transactions in Beijing and the fee is split equally between sellers and buyers."2
This structural policy shift proved more impactful than a one-time penalty. While regulators left the platform's operational architecture intact, they constrained its pricing power. Because ACN governs how commissions are split among participating agents without controlling total commission rates, the company possessed limited defensive leverage against mandated rate cuts.
Shock two: the short seller. On December 16, 2021, Muddy Waters Capital announced a short position and published a research report alleging that KE Holdings had systematically inflated its active agent headcount, store counts, gross transaction value, and net revenues.2 The company refuted the allegations, stating the report contained material factual errors.8 The following day, the board's audit committee initiated an internal review conducted by independent legal counsel and forensic accountants from a Big Four firm other than its primary auditor. On January 28, 2022, the company announced that the review was substantially complete and had failed to substantiate the short seller's claims.2
Shareholder class-action litigation followed on December 30, 2021, in the U.S. District Court for the Southern District of New York (Chin v. KE Holdings Inc.), alleging overstated operating metrics.2 On February 26, 2024, the court granted in part and denied in part the company's motion to dismiss. Rather than repleading the fraud claims, the plaintiffs amended their complaint to eliminate all Exchange Act claims, pursuing only Securities Act claims concerning the accuracy of total store and agent counts disclosed as of September 30, 2020, in the November 2020 follow-on offering prospectuses. Individual officer and director defendants were voluntarily dismissed in October 2024. On September 8, 2025, the parties executed a settlement agreement resolving the action for $4.95 million with no admission of liability, receiving final judicial approval on February 27, 2026.2
Evaluated objectively, the financial scale of the settlement was minor—representing a fraction of the $921 million in stock the company repurchased in FY2025 alone.1 The surviving claims were technical, and the fraud allegations were abandoned before trial. While the outcome did not confirm the broad fabrication alleged by short sellers, it highlighted an ongoing vulnerability: store and agent counts remain self-reported operational metrics lacking third-party auditing. Following the litigation, management shifted its primary operational disclosures away from headcounts toward productivity metrics.
Shock three: the one that actually mattered. The default cascade initiated by Evergrande (恒大集团) in late 2021 quickly spread to Country Garden (碧桂园), Sunac (融创中国), and the broader Chinese development sector. For Beike, developer distress went far beyond market sentiment. New home transaction services—collecting sales commissions from property developers—had served as the company's primary growth engine, expanding from RMB 7.5 billion in revenue in 2018 to RMB 46.5 billion in 2021.9 Those receivables were owed by the exact development entities entering financial restructuring.
The resulting contraction severely impacted financial performance. New home transaction revenue declined from RMB 46.47 billion in 2021 to RMB 28.65 billion in 2022. Existing home revenue fell from RMB 31.95 billion to RMB 24.12 billion. Total net revenues contracted 25% in a single year, falling from RMB 80.75 billion to RMB 60.67 billion, resulting in a net loss of RMB 1.397 billion.9 Annual platform GTV dropped from RMB 3,853.5 billion in 2021 to RMB 2,609.6 billion in 2022, while total transaction volume fell from over 4.5 million deals to approximately 3.8 million.9
The platform's physical infrastructure contracted alongside transaction volumes: active agents fell from over 406,000 at year-end 2021 to over 349,000 a year later, while active stores declined from over 45,000 to approximately 37,400 across 268 brands.9 This operational contraction provided empirical evidence regarding network retention: when transaction volumes dropped, approximately one-seventh of the active agent base and one-sixth of storefronts departed the network. Because partner agents operate as independent economic participants rather than locked-in users, network density fluctuates directly with transaction velocity.
In response to heightening geopolitical and regulatory uncertainty, KE Holdings executed a dual-primary listing on the Hong Kong Stock Exchange in May 2022, trading under ticker 2423 starting May 11.1617 The listing served primarily as a defensive strategy against potential delisting risks under the U.S. Holding Foreign Companies Accountable Act, ensuring the company maintained access to public capital markets even if its U.S. ADR line faced disruption. The dual-primary listing raised no net new capital and occurred near the trough of the domestic property cycle, securing institutional flexibility during an operational downturn.
Having navigated the immediate downturn, management confronted a more fundamental question: what the platform's core transaction engine was genuinely worth on the other side of a structural real estate crash.
VII. The Core Engine: Existing- and New-Home Transactions — Industry Structure and Economics
Materiality disciplines any analysis of KE Holdings' revenue streams. In FY2025, existing-home transaction services generated RMB 25.02 billion in revenue, while new-home transaction services generated RMB 30.60 billion. Together, these core services produced RMB 55.60 billion, accounting for 58.8% of the company's RMB 94.58 billion total revenue.2 They also represent the platform's most profitable segments, generating an FY2025 contribution of RMB 9.8 billion for existing-home transactions and RMB 7.6 billion for new-home transactions—implying contribution margins of roughly 39% and 25%, respectively, compared with 31% for home renovation and under 9% for rental management.1
Every diversification initiative across the company's portfolio involves expanding into sectors that generate lower profit per unit of revenue than the core brokerage engine. That margin gap does not invalidate diversification, but it demands an accurate evaluation of unit economics.
How the market is actually structured. China's residential transaction market is vast and extraordinarily fragmented by developed-market standards. At its 2021 peak, the platform cleared RMB 3.85 trillion in GTV—a transaction scale unmatched by any competitor.9 Named rivals occupy distinct layers of the market stack: 58同城 58.com and its property vertical 安居客 Anjuke, alongside 房天下 Fang.com, function primarily as listing portals monetizing lead generation; 我爱我家 I Love My Home operates as a traditional offline brokerage chain; and Century 21 China's franchise network integrated into Beike's platform rather than competing against it.
Listing portals have steadily surrendered ground to Beike due to a structural economic distinction. Portals sell advertising leads and lose visibility once a user exits the site. Beike embeds itself directly within the transaction flow—governing listing standards, commission splits, purchase contracts, escrow services, and deal closings. This integration allows the platform to capture verified transaction data rather than top-of-funnel browsing activity, while embedding its software into agents' daily operational workflows. Lead generation remains a media model with media-level margins; transaction clearing functions as essential market infrastructure.
How Beike wins — evaluated as evidence rather than assertion. Analysts frequently highlight four competitive mechanisms. The Housing Dictionary's census of 303 million properties represents the strongest moat, given the prohibitive cost and time required to replicate the physical survey.2 The Agent Cooperation Network's commission framework is effective yet conditional. Escrow and closing services provide a meaningful scale advantage, though not a monopoly lock. Meanwhile, Lianjia's directly operated network—comprising approximately 4,900 active stores and over 90,000 active agents—acts as a quality anchor that establishes operational standards across the broader platform.2
The most vulnerable claim is ACN's durability under take-rate pressure, as demonstrated by recent regulatory actions. In September 2023, the platform reduced existing-home commission rates in Beijing following government guidance urging equal fee-sharing between buyers and sellers.2 While a cooperation protocol that systematically splits fees remains a valuable operational asset, it does not confer independent pricing power. Any investment thesis premised on ACN's ability to arbitrarily raise take rates is contradicted by management's executed compliance.
The 2025 stress test — and its 2026 sequel. In 2025, management conducted a major network expansion that was subsequently reversed within twelve months.
Facing persistent market weakness in FY2025, the company expanded its physical footprint aggressively. Active store count rose 18.5% year over year to 61,139 and agent headcount increased 4.6% to 523,009, even as total platform GTV fell 5.0% to RMB 3,183.3 billion.1 Positioned at the time as a counter-cyclical market-share capture strategy, the financial outcomes proved disappointing. Gross margin compressed from 24.6% to 21.4%, operating margin dropped from 4.0% to 2.2%, net income fell 26.7% to RMB 2.99 billion, and non-GAAP net income declined 30.4% to RMB 5.02 billion.1
Two additional disclosures from the FY2025 report further illustrate the cost of that expansion. Net cash used in operating activities turned negative at RMB 376 million, down sharply from positive operating cash flows of RMB 9.45 billion in 2024 and RMB 11.41 billion in 2023.2 Additionally, the company recognized RMB 116.3 million in goodwill impairment across existing- and new-home reporting units in select cities, citing deteriorating local market conditions.2 A year of aggressive footprint growth yielded margin compression, a three-year profit low, negative operating cash flow, and asset write-downs on regional acquisitions.
Management pivoted quickly in early 2026. The operational pullback was evident within a single quarter: by March 31, 2026, total storefronts stood at 60,383—up 6.2% year over year against a low baseline, but below the December 2025 peak—while total agents fell 4.2% to 526,945 and active agents dropped 7.6% to 453,438.10 Shedding agents faster than physical stores indicated a concerted shift toward agent productivity over network size.
By the first quarter of 2026, fixed labor costs in the existing-home segment had fallen 24% year over year. CFO Xu Tao cited work on Lianjia—specifically "refining store and agent scale, optimizing organizational structure"—as the primary driver of margin expansion.11 By the second quarter of 2026, storefronts dropped 0.4% year over year to 60,274, total agents fell 3.1% to 540,634, and active agents decreased 7.5%.12
The empirical evidence supports a nuanced conclusion. The 2025 expansion was not a disciplined, market-share compounding initiative; it was an investment that executive leadership unwound within a year. The earnings recovery in 2026 resulted primarily from eliminating those added operational costs. While this reflects agile adaptive management, it disproves the premise that Beike can profitably expand its footprint through an industry downturn.
Beneath the cost reductions lay a distinct improvement in unit productivity. On the second-quarter 2026 earnings call, CEO Peng Yongdong noted that performance gains stemmed "not from expanding our network or rising prices," but from operational efficiency: existing-home transaction volume grew nearly 25% year over year on a stable store base, connected-store volume rose nearly 30%, and average transactions per active connected store climbed 26%. Simultaneously, existing-home contribution margin expanded 6.1 percentage points to 46.1%.1312 Simultaneous growth in volume and margin on a flat physical footprint highlights genuine operating leverage rather than pricing or product-mix shifts.
Why GTV and revenue have decoupled. The divergence between transaction volume and top-line growth stems from an accounting distinction that management highlights quarterly. When a connected non-Lianjia agent completes a transaction, Beike collects a platform service fee and records revenue on a net basis. When a Lianjia agent closes a deal, the company records the full gross commission. Consequently, as third-party connected agents handle a larger share of total volume—fulfilling the platform's core strategic objective—reported revenue grows more slowly than GTV while operating margins expand. In the second quarter of 2026, non-Lianjia platform service revenue grew 27.8% year over year, outstripping non-Lianjia GTV growth.13 Revenue declines accompanying GTV growth reflect a favorable margin mix rather than operational weakening.
Developer credit — the risk that actually bites. New-home transaction commissions represent receivables from property developers, exposing Beike to widespread real estate default risks. In response, management aggressively tightened credit terms. Accounts receivable turnover for new-home services dropped from 105 days in 2022 to 64 days in the first quarter of 2026, and reached approximately 39 days in the second quarter of 2026—a 12-day year-over-year reduction.91113 Management prioritized corporate-level agreements "with selected developers, including state-owned developers, to improve the quality of our sales projects and certainty of commission collection."2
Despite these collections, credit risk remains present. As of December 31, 2025, gross new-home accounts receivable totaled RMB 3.56 billion against a credit loss allowance of RMB 1.20 billion—representing a reserve ratio of roughly one-third.2 In the second quarter of 2026, the company recorded an additional bad debt provision of approximately RMB 280 million following what the CFO characterized as a prudent re-evaluation of receivables and collateral valuations, driving an 18.9% sequential increase in general and administrative expenses.13 Credit exposure is being actively managed, but not eliminated.
Myth versus reality. Three prevailing market assumptions regarding Beike fail under empirical analysis. First, framing Beike as the "Zillow of China" ignores basic business models: Zillow sells advertising leads, whereas Beike executes end-to-end transactions, resulting in a higher take rate, heavier cost structure, and deeper regulatory oversight. Second, claims of Tencent controlling the enterprise are inaccurate; Tencent held 8.7% of economic equity and 6.4% of aggregate voting power as of February 2026—far below the voting control held by executive leadership through Class B shares.2 Third, interpreting revenue declines as operational decay overlooks accounting mechanics; revenue contractions in existing-home platform services and Carefree Rent primarily reflect a structural shift toward higher-margin, net-basis fee recognition.
Porter, briefly and specifically. Applying Competitive Forces analysis reveals a distinct structural profile. Threat of new entry remains low due to the immense lead time and capital required to build the Housing Dictionary and store network. Supplier power resides in real estate agents, who demonstrated in 2022 that they exit the platform when market volumes contract. Buyer power is bifurcated between major property developers, who dictate receivable terms, and municipal governments, whose policy interventions dictate transaction volumes. Substitutes such as digital portals and direct seller-to-buyer models exist but face persistent structural disadvantages. While internal rivalry is softened by ACN converting competitors into network participants, the dominant competitive force remains state regulation, which retains direct authority over transaction pricing.
The demographic ceiling. Beyond cyclical fluctuations, long-term industry growth faces structural demographic constraints. Declining birth rates and a stabilizing urban population reduce long-term household formation, establishing a natural cap on transaction volumes. Cyclical policy stimulus alters short-term activity levels without reversing that demographic trend. CEO Peng's first-quarter 2026 remarks implicitly acknowledged this ceiling, shifting the strategic focus "from organizing supply to delivering decision support"—a strategy focused on increasing value capture per transaction rather than relying on volume expansion.11
Why win, why not. The core transaction business succeeds if the structural transition toward existing-home sales persists—management noted that resale transactions exceeded 50% of total national residential sales area in the first half of 202613—and if second-quarter 2026 connected-store productivity gains can be replicated nationwide rather than remaining confined to major metropolitan centers. Conversely, the segment faces downside risks if regulatory authorities expand the 2023 Beijing commission model nationwide, eroding take rates faster than productivity can offset, or if transaction volumes stall, forcing another cycle of footprint expansion and retrenchment.
Addressing these structural risks has required expanding beyond a single core market—leading directly into the company's newer business initiatives.
VIII. Diversification I: Home Renovation and the Shengdu Acquisition
The strategic logic of home renovation is close to self-evident, which is exactly why it deserves the harder look.
A brokerage possesses something no renovation company can buy: the customer relationship at the precise moment a home changes hands. The buyer of a resale apartment in Beijing is, with high probability, about to spend a meaningful sum making it habitable to their taste. Beike knows who they are, what they bought, the unit's floor plan — the Housing Dictionary again — and when they take possession. Customer acquisition cost in renovation is the industry's dominant expense. Beike's is close to zero.
The company launched its own renovation brand, Beiwoo, in April 2020, then bought scale.2 The Shengdu acquisition was announced in July 2021 with consideration capped at RMB 8 billion and closed in April 2022 at RMB 3.92 billion in cash plus 44,315,854 Class A ordinary shares.215
Four years of subsequent evidence let us test the two competing readings of that price cut.
The segment grew impressively at first — from RMB 5.05 billion of revenue in 2022 to RMB 10.85 billion in 2023 and RMB 14.77 billion in 2024, reaching RMB 15.43 billion in FY2025, up 4.4%, with contribution margin around 31%.921 By that measure it is exactly what management promised: a real, profitable second leg.
Then it went into reverse. Q1 2026 renovation revenue fell 20.6% year over year to RMB 2.3 billion; Q2 2026 fell 30.1% to RMB 3.19 billion.1112 Management attributes this to deliberate choices — exiting inefficient customer acquisition channels and withdrawing from cities with weak unit economics — compounded by genuinely weaker demand, since fewer new home deliveries mean fewer renovation projects and, as the CFO put it, companies that previously focused on new homes are "flooding into the existing home market, intensifying the competition."13
Contribution margin, meanwhile, went the other way: 36.2% in Q1 2026 and 39.6% in Q2, driven by centralized procurement that lowered some material prices by more than 20%, plus a shift to platform-selected and platform-dispatched tradespeople whose monthly order volumes rose over 50% versus the second-half-2025 average.1113
So how should the acquisition be judged? On price, the renegotiation looks vindicated — a business now running below its 2024 revenue would plainly not have been worth RMB 8 billion. On thesis, the picture is less flattering. The strategic argument was that owning the customer at the point of purchase would produce durable, compounding renovation demand. Four years on, the segment's revenue is contracting at 20–30% annually and management has stopped defending scale, telling analysts explicitly that "we will not trade profitability for scale" and that contract value, not revenue, is the leading indicator to watch.13
The calibrated conclusion: the deal was bought at a defensible price and has been managed into genuine profitability, but the growth thesis that justified making it a strategic pillar is not currently supported by the numbers. Renovation today is a well-run, margin-improving, shrinking business — a profit contributor, not a growth engine. The falsifiable test going forward is whether contract value inflects before revenue does; if front-end contract value stays weak into 2027, the strategic-adjacency argument should be retired regardless of margin.
The second wing has the opposite problem — and a much more interesting one.
IX. Diversification II: Home Rental Services and "Carefree Rent" (省心租)
If home renovation represents a profitable business that stopped growing, 省心租 Carefree Rent presents the opposite phenomenon: an expanding operation whose top-line revenue is contracting primarily due to accounting shifts.
The operational trajectory represents one of the company's most rapid expansions. Home rental services revenue expanded from RMB 6.10 billion in 2023 to RMB 14.33 billion in 2024, reaching RMB 21.90 billion in FY2025—a 52.8% year-over-year increase.21 By FY2025, the segment accounted for 23.2% of total net revenues, up from under 8% two years prior.2 That same year, the business achieved full-year profitability for the first time, generating a contribution margin of nearly 8.6%.1
However, top-line contraction followed: revenue declined 1.5% year over year in the first quarter of 2026 and fell 14.8% in the second quarter to RMB 4.83 billion.1112
This revenue decline reflects a fundamental change in business structure rather than operational weakening. In its original form, Carefree Rent functioned primarily as a master-lease model: Beike leased properties directly from owners with guaranteed rental payouts, then subleased them to tenants while absorbing vacancy risk. Accounting rules required recognizing the full tenant rent as gross revenue and owner payouts as operating costs, producing high top-line figures alongside thin margins and vacancy exposure. Management has since pivoted toward a decentralized, agency-based property management model. Operating as a manager rather than a principal, the platform collects a management fee and recognizes revenue on a net basis—reducing reported top-line revenue by roughly four-fifths while generating higher contribution margins and eliminating balance-sheet vacancy risk.
Unit volume and margin expansion reveal underlying operational growth. Managed units surpassed 740,000 by the end of the first quarter of 2026—a 47% year-over-year increase—and crossed 790,000 by the end of the second quarter, up 34%, with net-basis products accounting for over 40% of managed inventory in March and exceeding 50% by June.1113 Contribution margin expanded from 6.7% in the first quarter of 2025 to 14.8% in the first quarter of 2026—marking six consecutive quarters of sequential improvement—and reached 15.3% in the second quarter of 2026, up 6.9 percentage points year over year.1112
Consequently, evaluating the segment strictly by top-line revenue reflects accounting methodology rather than operational health.
The durability of these margin gains depends on operational efficiency metrics disclosed by management. Managed units per asset manager increased approximately 40% year over year to roughly 170 units. Property owner renewal rates reached 74%—a 4-percentage-point increase—while tenant renewal rates rose 1 percentage point to 56%.13 CFO Xu Tao highlighted unit renewal rates as the primary determinant of long-term profitability, noting that sustained margins require "managing an asset pool with a lower churn, fewer re-leases, and higher renewals" to ensure labor and channel acquisition costs grow more slowly than revenue.13 In property management, the expenses associated with re-leasing units—including vacancy downtime, marketing fees, and agent labor—directly dictate portfolio returns. An owner renewal rate of 74% implies that approximately one-quarter of the managed portfolio must be re-contracted each year.
Management highlighted the strategic importance of the rental segment prior to its profitability. When revenue growth across core transaction services decelerated to low single digits, management directed market analysts toward the rental portfolio as a counter-cyclical growth driver.24 Operational unit volume and margin expansion have validated that strategy, even as net-basis revenue reporting has reduced headline segment revenue.
Strategically, property management provides a structural hedge driven by demographic shifts rather than cyclical trends. Homeownership rates among younger urban residents in China have declined, making rental demand substantially less sensitive to transaction cycles than residential sales. As existing- and new-home sales volumes fluctuate, a recurring portfolio of 790,000 managed units generates steady cash flow throughout macroeconomic cycles.
However, several structural limitations temper this growth potential. First, segment economics remain lower than core brokerage operations: a 15.3% contribution margin in rental management compared with 46.1% in existing-home transactions means that each renminbi of rental revenue yields approximately one-third the gross profit of transaction revenue.12 Consequently, rental volume provides a revenue buffer rather than a full profit replacement. Second, management cautioned against assuming linear margin expansion, noting that quarterly margins "may fluctuate" depending on product mix and operational execution.11 Third, potential governance conflicts persist: 自如 Ziroom, a rental management platform whose principal shareholder is the founder's widow and which shares a board director with KE Holdings, offers leasing services that company filings acknowledge "may potentially compete with our rental property management services, including Carefree Rent."9 This operational overlap between a related entity controlled by the principal shareholder group and the company's fastest-growing segment represents an unaddressed governance friction.
The key strategic question is whether contribution margins continue expanding as net-basis units exceed 50% of the portfolio, or whether they stabilize in the mid-teens. The outcome will determine whether home rental services matures into a high-return profit engine or serves as low-margin operational ballast.
Beyond home renovation and property management lies a third expansion initiative operating under a fundamentally different model.
X. The New, Riskier Bet: Beihome and Residential Development
The central premise of the company's valuation has long been its asset-light platform model. It does not own physical housing stock, hold land reserves, or assume construction risks; instead, it clears transactions between buyers and sellers in exchange for platform fees.
In 2023, management expanded its corporate strategy from "one body, two wings" to "one body, three wings" (一体三翼), introducing the Beihaojia unit—or Beihome—as a third pillar "to facilitate supply-side upgrades for new homes."2 In operational terms, this marked KE Holdings' entry into direct land acquisition.
Examining the exact financial scale of this venture provides necessary context. As of December 31, 2025, the company directly managed two residential land parcels in Chengdu and Shanghai, both acquired in 2024.2 Real estate properties under development on the balance sheet doubled to RMB 2.23 billion, up from RMB 1.09 billion a year earlier. Within that total, land-use rights accounted for RMB 1.8 billion, up from RMB 1.1 billion. Crucially, the entire RMB 1.8 billion in land-use rights was pledged as collateral for project-specific bank loans as of December 31, 2025, compared with zero pledged assets a year prior.2
For a company carrying approximately RMB 67.3 billion in broad cash, RMB 2.23 billion in development inventory does not present an immediate solvency risk.13 However, it signals a distinct strategic shift: moving from unencumbered operations to pledging RMB 1.8 billion in land assets within a single year.
Management frames this initiative around a customer-to-manufacturer (C2M) model. Under this rationale, Beike's proprietary transaction data provides granular insights into buyer preferences, enabling developers to avoid building mismatched housing inventory. According to its FY2025 annual report, Beihaojia provided third-party developers with data-driven decision support spanning land acquisition assessment, product positioning, pricing, and marketing across several collaborative projects, framing its two directly held parcels as pilot vehicles to validate these analytical capabilities.2
While monetizing market intelligence while holding select pilot parcels presents a logical thesis, the operational framework is more complex. Related-party disclosures reveal an ecosystem of "Beihaojia business equity investees" over which KE Holdings exercises significant influence. This network generated RMB 74 million in agency service revenue for the company in 2025. At year-end, KE Holdings held RMB 289 million in short-term loan receivables and RMB 16 million in long-term loan receivables from these investees, while simultaneously carrying RMB 498 million in short-term and RMB 259 million in long-term loans payable to them.2 Capital is moving bidirectionally between the company and unconsolidated entities in which it holds partial stakes.
This dual capital flow presents a clear analytical challenge. The core transaction business suffered severe losses when property developers defaulted on obligations, prompting management to spend five years shortening receivable turnover from 105 days to 39 days while provisioning a one-third credit loss allowance against new-home receivables.213 Yet alongside those credit-tightening measures, the company has entered residential development directly—assuming balance-sheet exposure to the very sector risks it sought to insulate against.
Management maintains that C2M remains fundamentally a data service rather than a construction business, a distinction supported by the modest current scale of two parcels. However, corporate disclosures offer explicit caveats. In its annual report, the company cautions that the model "is at an early stage and subject to changes" and warns that property investments under Beihaojia "may not yield the anticipated returns, or any returns at all," potentially leading to "the recognition of significant impairment charges."2 Such explicit risk disclosures indicate that execution challenges remain substantial.
A notable divergence exists between how the third wing is marketed and how it is funded. In public presentations of the "one body, three wings" strategy, executive messaging has focused primarily on artificial intelligence tools integrated across the platform, rather than on land acquisition or property development economics.22 Consequently, a business unit that absorbs balance-sheet capital is communicated largely as a software and analytics initiative.
This narrative positioning is compounded by limited reporting transparency. KE Holdings does not report Beihaojia as an independent operating segment, nor does it disclose committed capital allocations or expected returns. Furthermore, management made no mention of the unit during its first-quarter or second-quarter 2026 earnings calls, despite detailed analyst inquiries regarding all other operating segments.1113 An initiative that doubled development inventory within a single year and encumbered RMB 1.8 billion in assets with project debt remained absent from quarterly executive commentary.
At present, Beihaojia remains too small to compromise group solvency, yet too opaque to evaluate fully. The assertion that C2M offers strategic value to core transactions remains unverified: disclosed figures demonstrate neither enhanced developer project returns nor capital impairment. However, capital deployment into the unit is expanding even as executive commentary excludes it from investor updates. The primary metric for tracking this venture remains the quarterly inventory balance: if real estate properties under development continue to expand significantly, characterizing the initiative as a minor pilot will become increasingly difficult to sustain.
XI. Current Management: Peng Yongdong's Capital Allocation and Credibility Record
Who controls the company. As of February 28, 2026, corporate ownership is significantly more concentrated than the public float implies. Entities affiliated with the late founder's widow held 849.6 million Class A shares, representing 25.1% of total shares outstanding. However, those shares carry an irrevocable proxy, with Propitious Global Holdings Limited authorizing Baihui Partners L.P. to exercise all voting rights.2 Chief Executive Officer Peng Yongdong personally held 5.1% of total economic equity and 22.5% of voting power. Together, Class B holders—Peng and co-founder Shan Yigang—controlled 29.9% of aggregate voting power through 138.6 million shares. Meanwhile, Tencent's investment vehicles held 295.2 million Class A shares, representing 8.7% of economic equity and 6.4% of voting power.2
This voting hierarchy reframes common market assumptions. The founding family holds the largest economic position but has contractually surrendered its voting authority to a management partnership. Strategic control rests with a small executive team holding a fraction of economic equity. Tencent, often described as a dominant backer, functions as a single-digit minority holder with less voting power than economic equity. Consequently, public shareholders possess virtually no mechanism to influence corporate governance or strategic decisions.
Executive cash compensation remains modest relative to corporate revenue, with KE Holdings paying an aggregate RMB 29.7 million in cash to all directors and executive officers combined in FY2025.2 Alignment relies primarily on equity grants, though share-based compensation expenses remained substantial—running RMB 3.22 billion in 2023, RMB 2.73 billion in 2024, and RMB 1.91 billion in 2025.2
The capital return record, told through outcomes. The company's share repurchase program expanded steadily alongside market downturns. The board authorized up to $1 billion across May and August 2022, increasing the cap to $2 billion in August 2023, $3 billion in August 2024, and $5 billion in August 2025, while extending the program term to August 31, 2028.21819 Execution accelerated accordingly: management repurchased approximately $921 million of stock in FY2025—up 29% year over year—followed by $195 million in the first quarter of 2026, and $250 million in the second quarter, which included the platform's initial buybacks in Hong Kong.11113 Cumulative repurchases reached approximately $2.99 billion through the second quarter of 2026, retiring roughly 14.8% of pre-program shares outstanding.13
Retiring nearly one-seventh of total equity during a prolonged real estate contraction—without relying on balance-sheet leverage—represents a disciplined execution of shareholder capital returns.
However, dividend distributions tell a more nuanced story. The board declared a special dividend of approximately $0.2 billion in August 2023, followed by final dividends of roughly $0.4 billion in March 2024 and March 2025. In March 2026, the board declared a reduced final dividend of approximately $0.3 billion.2 This 25% cut reflected FY2025's negative operating cash flow, demonstrating financial flexibility rather than fixed payouts. While management acknowledged the reduction openly, claims of an uninterrupted dividend expansion across the downturn are inaccurate; cash returns contracted in the same fiscal year that the company expanded development land holdings and pledged assets against project debt.
The narrative test across four regimes. Evaluating executive consistency requires examining how management framed strategy across changing market cycles.
During the FY2022 loss year, executive commentary prioritized operational survival—emphasizing cost reductions, developer credit controls, and platform resilience. In FY2023, as net income rebounded to RMB 5.89 billion, messaging shifted toward portfolio diversification, framing home rental and renovation as engines designed to decouple revenue from property cycles.223 By FY2024 and FY2025, the strategy pivoted to market-share expansion, aggressively adding stores and agents into a weakening housing market. From the first quarter of 2026 forward, management replaced footprint expansion with strict operational efficiency.
This progression reflects tactical flexibility, yet it highlights an operational reversal. The 2024–2025 expansion and the 2026 efficiency initiative represented opposing strategic directions executed within eighteen months, requiring leadership to dismantle the footprint added during the previous expansion. CEO Peng addressed the pivot directly during the second-quarter 2026 call, stating that even if transaction volumes recover, the company "will not return to extensive expansion" and confirming that new capital deployments "must pass stage-gated ROI and service validations before scaling."13 While executive candor clarifies current priorities, it also confirms the costs incurred during the prior expansion phase.
Furthermore, annual earnings performance reveals persistent underlying pressure. Net income attributable to shareholders fell from RMB 5.89 billion in 2023 to RMB 4.08 billion in 2024, and RMB 2.99 billion in 2025—marking three consecutive annual declines.21 Framing the early 2026 efficiency gains as a permanent profit turnaround relies on two quarters of operating data against three years of lower net income.
Nonetheless, those two quarters showed genuine margin expansion. In the first quarter of 2026, non-GAAP operating profit rose 45.1% year over year to RMB 1.67 billion despite a 19% drop in revenue, as total GAAP operating expenses fell 22.3% to a three-year low.11 In the second quarter, non-GAAP net income reached RMB 3.19 billion—up 74.9% year over year—yielding a non-GAAP net margin of 13.0%, while operating cash flow swung positive to RMB 6.61 billion.1213 CFO Xu Tao characterized these results as "a structural improvement rather than a cyclical one."11 Simultaneous margin expansion across all operating segments supports that assessment, though a 6.3% year-over-year recovery in second-quarter platform GTV indicates that broader market volume also contributed to the rebound.
The AI question, translated. Executive updates throughout 2026 emphasized artificial intelligence initiatives. Concrete applications remain focused on agent productivity. Disclosures highlight an internal tool-building platform where employees use natural-language prompts to generate custom software workflows, engaging over 7,100 staff members and deploying more than 4,400 active applications by late April 2026. Additionally, a specialized client-manager system, compensated independently of transaction closings, handled over 50,000 leads between May and July 2026, achieving a 7.4% lead-to-showing conversion rate compared to an industry baseline near 5%.1113 Management contends that AI tools automate standardized real estate information while enhancing agent advisory capabilities.
This strategic framing presents a clear rationale: it positions a network of 523,009 agents and a database of 303 million properties as direct beneficiaries of automation. The core metric for evaluating this strategy remains agent productivity. In the first quarter of 2026, Lianjia's per-capita transaction volume rose 26% year over year, while per-capita commission revenue grew 8.5%.11 Sustained productivity growth on a stable headcount would demonstrate software efficiency, whereas plateauing output would indicate that AI tools function primarily as cost-management narrative support.
Trust flags, weighed. Several operational features warrant ongoing monitoring. The Muddy Waters short-seller litigation concluded with a $4.95 million settlement over narrowed disclosure claims with no admission of wrongdoing—a financial outcome too small to validate the fraud allegations, yet one that leaves self-reported store and agent metrics unverified by independent auditors.2 Potential governance overlaps persist in the rental segment, where Ziroom operates as a related party alongside Carefree Rent.9 Additionally, capital deployment into the Beihaojia development unit continues without separate segment reporting. Finally, the FY2025 sequence—combining store expansion, margin compression, negative operating cash flow, goodwill write-downs, a reduced dividend, and land asset pledges—underscores the need for disciplined oversight.
In summary, leadership has demonstrated strong crisis execution, maintained substantial share repurchases during an industry contraction, and delivered meaningful operating efficiency in early 2026. However, strategic narrative shifts over the past eighteen months incurred real capital costs, three-year net income trends show continuous decline, and capital allocation toward property development lacks reporting transparency. Current management remains an effective operational team navigating an unproven capital-allocation expansion.
XII. Bull vs. Bear: The Investment Case
The bull case. Beike holds a dominant position in China's residential housing market that competitors have failed to replicate. Its Housing Dictionary, covering approximately 303 million properties, represents seventeen years of physical fieldwork that cannot easily be duplicated.2 Meanwhile, the Agent Cooperation Network has converted 279 rival brokerage brands into platform participants, with non-Lianjia agents generating a substantial majority of platform transaction volume—confirming the strategic success of the open-platform model.2
Operating results from 2026 provide compelling support for the bull thesis because margin gains were driven by structural efficiency rather than market-wide volume recovery. Contribution margins expanded simultaneously across all major business segments: existing-home transactions reached 46.1%, new-home transactions hit 28.8%, home renovation rose to 39.6%, and home rental services increased to 15.3%—with each line improving both year over year and quarter over quarter.12 Average transactions per connected storefront climbed 26% on a flat physical footprint,13 while operating cash flow swung decisively back into positive territory.
Diversification has become a tangible revenue contributor. Together, home renovation and rental services generated RMB 37.3 billion in revenue during fiscal year 2025,2 accounting for 39.5% of total top-line revenue as rental achieved full-year profitability,1 with its contribution margin nearly doubling within a year.12 Financial liquidity remains robust: the company ended 2025 with approximately RMB 67.3 billion in broad cash13 against just RMB 208 million in short-term borrowings,2 leaving the platform virtually debt-free in a heavily leveraged sector. Management has complemented this liquidity with consistent shareholder returns, repurchasing and retiring 14.8% of outstanding shares since September 2022.13
Framed through competitive strategy, Beike relies on three primary moats: a cornered resource in the Housing Dictionary, network economies in the ACN platform, and expanding process power. While data assets and network effects attract attention, process power—the accumulated operational know-how required to run a high-density, multi-brand brokerage network—may prove most durable. The 2026 per-store productivity gains stem directly from these refined operational workflows, an advantage that requires years to build and is equally slow to erode.
The bear case. The core residential market faces long-term structural and cyclical pressures. Cyclically, gross transaction value has failed to regain its 2021 peak of RMB 3.85 trillion,9 ending 2025 down 5.0% year over year at RMB 3.18 trillion.1 Structurally, declining birth rates and a stabilizing urban population limit long-term household formation, establishing a permanent demographic ceiling on transaction volumes.
Take-rate vulnerability represents a major unpriced risk. The September 2023 commission reduction in Beijing, implemented following regulatory guidance on fee sharing, demonstrated that transaction fees remain subject to policy intervention.2 While Beike's network scale secures transaction volume, it does not confer independent pricing power—and the platform's primary test of fee defense resulted in lower commissions.
The company's 2025 performance illustrates these structural constraints. An 18.5% expansion in store footprint yielded gross margin compression from 24.6% to 21.4%, operating margin contraction from 4.0% to 2.2%, and a 26.7% decline in net income.1 The expansion also generated negative operating cash flow, triggered goodwill impairments, and forced management to reverse the expansion within a year.2 An intermediary that cannot expand its physical footprint profitably during a market slowdown possesses a narrower competitive moat than optimistic valuations assume.
Furthermore, the quality of business diversification remains mixed. Home renovation revenue is contracting between 20% and 30% year over year.12 Meanwhile, rental services revenue has declined due to net-basis accounting shifts, and its contribution margin—despite recent gains—remains roughly one-third that of existing-home transactions.12 While both initiatives contribute positive operating profit, neither currently functions as a top-line growth driver.
The Beihaojia development unit presents the clearest strategic contradiction. A company valued on asset-light platform economics has entered direct land acquisition, doubling development inventory to RMB 2.23 billion in a single year while pledging RMB 1.8 billion in land-use rights as loan collateral.2 This shifts balance-sheet capital directly into the real estate development sector that previously caused severe credit losses in the core business. Compounding this risk, the unit lacks separate financial segment reporting and received no mention during executive commentary on consecutive 2026 quarterly earnings calls.
Additional governance and structural risks persist across the corporate structure: insider leadership controls 29.9% of aggregate voting power despite holding a fraction of economic equity, domestic operating assets remain bound by variable interest entity agreements, Ziroom operates as a related party competing in rental management, self-reported store and agent metrics remain unverified by third-party audits, and dual-listed Chinese equities face ongoing geopolitical regulatory overhangs.29
Where the evidence lands. Empirical evidence validates a narrowed version of the competitive moat. The historical record—including agent departures in 2022, commission cuts in 2023, and the unwound store expansion in 2025—refutes claims of unassailable pricing power or counter-cyclical growth. However, it confirms a valuable distribution and data asset that enables the company to preserve or expand volume share in a contracting market, protecting margins through operational cost controls rather than price increases. That foundation represents a resilient operational business, though not a high-compounding growth platform.
Executive leadership has proven effective in crisis management and operational streamlining, but its long-term capital allocation track record remains unproven. Diversification into home renovation and rental management has established secondary profit streams, though neither segment currently functions as a top-line growth driver. Meanwhile, the Beihaojia property development venture remains under-disclosed and unvalidated.
Three KPIs, and only three.
First, existing-home GTV. Gross transaction value in existing homes provides the clearest indicator of housing market recovery, undistorted by net-revenue accounting shifts or developer credit risk. Segment GTV grew 8.0% year over year in the second quarter of 2026 following a 7.9% decline in the first quarter.1211 Sustained growth over subsequent quarters would signal a broader market stabilization, whereas a renewed decline would indicate that early 2026 gains were confined to major metropolitan centers.
Second, managed rental units multiplied by rental contribution margin. Tracking unit count alongside margin provides a true measure of operational growth, bypassing accounting distortions in top-line rental revenue. Managed inventory reached 790,000 units with a 15.3% contribution margin in the second quarter of 2026.12 Continued expansion across both metrics would confirm home rental services as a compounding profit engine; a margin plateau in the mid-teens alongside slowing unit growth would characterize the segment as a volume hedge.
Third, real estate properties under development relative to shareholder returns. At year-end 2025, development inventory stood at RMB 2.23 billion,2 compared with approximately $1.2 billion in total stock repurchases and dividends for the fiscal year.1 This quarterly balance-sheet comparison serves as the primary metric for determining whether leadership maintains its capital-light discipline or shifts resources toward asset-heavy property development risks.
XIII. Business and Investing Lessons
Opening a moat can be the moat—but only after solving the problem others cannot. The Agent Cooperation Network (ACN) pivot is routinely cited as a lesson in platform generosity, but it was fundamentally a masterclass in operational sequencing. Lianjia spent seventeen years building the Housing Dictionary and establishing authentic listing standards before opening its system to rivals. Consequently, when competitors finally joined, they did so on Beike's terms rather than negotiating their own. A company that opens its platform before establishing an essential proprietary asset does not build a network; it simply surrenders product value. The generalizable rule is that platform economics beat proprietary lock-in only when the platform operator controls a genuinely scarce input—and that scarcity must be verified before opening the network, not assumed after the fact.
Surviving concentrated, simultaneous shocks is a data point—and the operational response is the real evidence. Between May 2021 and December 2022, KE Holdings lost its founder, faced regulatory inquiries, confronted short-seller allegations, and watched its core revenue line contract by half. Its response provides the empirical test: management shortened receivable turnover from 105 days to 39 days, disciplined store density by paring back agent headcount, expanded renovation and rental management from near zero to more than RMB 37 billion in combined annual revenue, and repurchased nearly 15% of outstanding shares. Many companies survive downturns passively by waiting out the market. The true test of resilience is whether the balance sheet and operating model emerge structurally stronger rather than merely intact.
The gravitational pull from asset-light to asset-heavy is a recurring failure mode for maturing platforms, and it invariably arrives framed as operational synergy. The thesis behind the Beihaojia development unit is logical on its face: Beike understands buyer preferences, developers often do not, and that data intelligence should command commercial value. Yet similar arguments have drawn asset-light platforms into inventory, logistics, manufacturing, and capital-intensive lending for decades. The risk is not that such ventures automatically fail, but that they consume capital and executive attention on a risk profile fundamentally different from the asset-light model investors originally underwrote. Furthermore, these initiatives are frequently disclosed with the least transparency during their fastest growth phase. Investor discipline here must be mechanical rather than narrative: track the exact balance-sheet line where capital accumulates every quarter, regardless of strategic positioning.
A related-party competitor operating in a platform's fastest-growing segment is not a routine footnote. The operational overlap between Ziroom and Carefree Rent is acknowledged in corporate disclosures and has persisted for years without structural resolution. Governance arrangements of this nature rarely produce a single identifiable violation; instead, they create persistent ambiguity over whether high-potential opportunities are allocated to the publicly listed company or to an entity owned by the controlling shareholder group. The broader rule applies: when a founder-family vehicle operates alongside a public company in an adjacent market, the appropriate analytical posture is not active suspicion of individual transactions, but a structural valuation discount applied to that segment's growth metrics.
Finally, read the underlying accounting policy before evaluating the headline growth rate. Two of Beike's core strategic evolutions—the net-basis fee recognition in existing-home platform services and the property-management transition in Carefree Rent—generated declining reported revenue alongside expanding operational volumes and improving profit margins. An investor screening the business solely on top-line revenue growth in 2026 would reach conclusions entirely opposite to the underlying operational reality.
XIV. Epilogue: What to Watch
As of early September 2026, KE Holdings occupies a distinct operational position: its profitability is expanding faster than its underlying market, its reported revenue declines reflect higher-margin structural shifts rather than operational contraction, and its primary long-term strategic uncertainty remains the initiative management discusses least.
The operating performance entering the second half of 2026 represents the company's strongest financial position since 2021. Gross transaction value returned to year-over-year growth in the second quarter, accompanied by contribution margin expansion across all core segments. Non-GAAP operating margin reached 14.6% and non-GAAP net margin reached 13.0%, both marking three-year highs.12 Single-quarter operating cash flow reached RMB 6.61 billion, more than reversing the cash flow deficit recorded across fiscal year 2025.13 In addition, the company initiated share buybacks in Hong Kong for the first time.
Executive guidance for the remainder of the year remains cautious. Chief Financial Officer Xu Tao warned analysts against extrapolating second-quarter profitability across future periods, noting that marketing incentives and frontline sales expenses will fluctuate based on revenue scale, segment mix, and seasonality.13 Chief Executive Officer Peng Yongdong framed strategy for the second half around factors within management's direct control: refining decision-support tools to attract customers while maintaining a disciplined cost structure to protect cash flow.13 By avoiding explicit numeric targets, leadership adopted a measured posture following the rapid footprint expansion and subsequent retrenchment of the previous eighteen months.
Three critical factors will determine the platform's trajectory over the next 12 to 24 months.
First, whether China's residential housing market has established a durable floor or if the early 2026 rebound was confined to top-tier cities. Peng noted that while prices in core metropolitan areas demonstrated resilience, "a broader nationwide stabilization will need a more month of data to confirm," observing that the housing market "remains polarized."1113
Second, whether recent agent productivity gains prove durable. Beyond cost reductions, operating results in 2026 were highlighted by a 26% year-over-year increase in average transactions per active connected store on a stable physical footprint.13 If these efficiency gains compound, the platform will demonstrate an ability to expand operating economics without expanding its physical store network—a dynamic that the 2025 footprint expansion failed to achieve.
Third, whether the Beihaojia real estate development unit remains a limited experiment. Maintaining two land parcels and RMB 2.23 billion in development inventory represents a pilot project. However, if development assets expand significantly on the balance sheet—particularly without separate financial segment reporting—it would signal a structural shift toward capital-intensive property development well before such a pivot is reflected in official strategic commentary.
The enterprise emerging from this real estate cycle operates under fundamentally different parameters than the company that completed its initial public offering in 2020. The original public thesis priced in expanding transaction volume across a growing housing market. The current business model focuses on operational efficiency and transaction quality within a smaller, highly regulated, and demographically constrained market. Whether this transition yields a smaller version of the traditional brokerage model or a more durable, higher-margin transaction infrastructure platform is the central question facing investors over the coming two years.
References
-
KE Holdings Inc. Announces Fourth Quarter and Fiscal Year 2025 Unaudited Financial Results and a Final Cash Dividend — GlobeNewswire, 2026-03-16 ↩↩↩↩↩↩↩↩↩↩↩↩↩
-
KE Holdings Inc. — Form 20-F, FY2025 Annual Report (SEC) ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩
-
SoftBank-Backed KE Raises $2.1 Billion in Above-Target IPO — Bloomberg, 2020-08-12 ↩↩
-
KE Holdings Inc. Announces with Profound Sadness the Passing of its Founder and Chairman ZUO Hui — Business Wire, 2021-05-20 ↩↩
-
Beike names CEO Peng Yongdong as new chairman following Zuo Hui's death — South China Morning Post, 2021-05-24 ↩↩
-
Zuo Hui, founder of China's biggest online real estate platform Beike Zhaofang, dies at 50 — South China Morning Post, 2021-05-21 ↩↩↩
-
Tencent's housing broker Beike is in regulators' cross hairs as China's antitrust investigations pick up pace, Reuters reports — South China Morning Post, 2021-05-25 ↩↩
-
KE Holdings Refutes the Muddy Waters' Report — Business Wire, 2021-12-17 ↩↩
-
KE Holdings Inc. — Form 20-F, FY2022 Annual Report (SEC), 2023-04 ↩↩↩↩↩↩↩↩↩↩↩
-
KE Holdings Inc. Announces First Quarter 2026 Unaudited Financial Results — Investors.ke.com, 2026-05-19 ↩
-
Transcript: KE Holdings Q1 2026 Earnings Conference Call — Benzinga, 2026-05 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩
-
KE Holdings Inc. Announces Second Quarter 2026 Unaudited Financial Results — GlobeNewswire, 2026-08-21 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩
-
KE Holdings (BEKE) Q2 2026 Earnings Call Transcript — The Motley Fool, 2026-08-21 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩
-
KE Holdings Inc. — Form 20-F, FY2021 Annual Report (SEC), 2022-04-19 ↩
-
KE Holdings Inc. Announces Closing of Shengdu Acquisition — KE Holdings IR ↩↩
-
KE Holdings Inc. Announces Proposed Dual Primary Listing on The Stock Exchange of Hong Kong Limited — KE Holdings IR, 2022-05-04 ↩
-
Online housing firm Beike prepares for dual-primary listing in Hong Kong — TechNode, 2022-05-06 ↩
-
KE Holdings Inc. — Form 6-K, buyback authorization increase to $3B (SEC), 2024-08 ↩
-
KE Holdings Inc. — Form 6-K, buyback authorization increase to $5B (SEC), 2025-08 ↩
-
This Property Broker Is Worth More Than China's Largest Developer — Yicai Global ↩
-
Beike Zhaofang Surges on NYSE Debut After Solid IPO Reception — South China Morning Post, 2020-08-13 ↩
-
Beike unveils new AI features under its 'one body three wings' policy — AIM Group, 2025-06-07 ↩
-
What property crisis? KE Holdings extends profit streak — Bamboo Works ↩
-
KE Holdings Looks To Rentals As Weak Chinese Market Caps Q3 Revenue Growth At 2% — Online Marketplaces ↩
LinkedIn
WhatsApp
Reddit
Hacker News
Email
Kindle
RSS Feed
Spotify
Apple Podcasts
Amazon Music
Audible
YouTube