QFIN

Core Business Description

Qfin Holdings (formerly Qifu Technology, formerly 360 DigiTech — the frequent name changes themselves are worth noting as a flag on brand continuity) operates China's largest independent Credit-Tech platform. The business sits between two enormous constituencies: hundreds of millions of Chinese consumers seeking credit, and banks and other licensed financial institutions that want to lend to them but lack the data infrastructure and distribution capability to do so efficiently at scale.

QFIN solves a structural information asymmetry in Chinese consumer credit. Traditional lenders struggle to underwrite thin-file borrowers — people with limited credit history who are nevertheless creditworthy. QFIN applies proprietary AI models, behavioral data analytics, and alternative data sources to score these borrowers, then connects them with institutional capital at a fee. The primary customer-facing product is the Qifu Jietiao mobile app (formerly 360 Jietiao), through which end consumers apply for personal loans typically used for consumption, education, and small business needs.

The company does not hold a banking license. It is fundamentally a technology-enabled middleman — a credit marketplace with sophisticated risk management bolted on top. This structure means it can scale without proportionally expanding its own balance sheet, but it also means its value proposition is perpetually under pressure from regulators who view fee extraction in lending intermediation with increasing skepticism.

Revenue Breakdown by Segment

Full-year 2025 total net revenue: RMB 19,205.1 million (~US$2.75 billion), up 11.9% from RMB 17,165.7 million in 2024. The apparent growth masks a deeply worrying sequential deceleration: Q3 2025 revenue was RMB 5,205.7M, while Q4 2025 saw a sharp contraction as the 24% interest rate cap began to bite. The revenue mix shift toward capital-heavy credit-driven services (from a peak of ~58% capital-light at year-end 2024 to a lower share by late 2025) runs counter to management's stated strategic intention of becoming more capital-light — a revealing contradiction.

Revenue Quality

Revenue quality is mixed. The repeat borrower contribution of 93.3% in FY2025 (up from 93.1% in 2024) indicates strong retention of active borrowers and suggests meaningful switching costs at the consumer level — a genuine positive. However, the underlying revenue is fundamentally transactional: it is earned as loans are originated and serviced, and a regulatory change to fee disclosure or rate caps can immediately restructure the economics. There are no multi-year contracts, no subscription revenues, and no software licensing agreements that provide baseline recurring income. The business requires constant loan origination volume to generate revenue — making it inherently cyclical and regulatory-sensitive.

Unit Economics

QFIN's unit economics look attractive when measured at the borrower level: 93%+ repeat borrower rates imply low marginal acquisition cost for the existing base. The company reports 38.9 million cumulative borrowers with drawdowns, 63.6 million users with approved credit lines, and 291.3 million registered users as of December 2025 — implying a conversion funnel of roughly 13% from registered users to active borrowers. Customer Acquisition Cost (CAC) is not publicly disclosed as a standalone metric, but can be inferred as modest given the funneling from Qifu Jietiao's app ecosystem. Lifetime value is driven by loan tenors averaging ~11 months and repeated cycling. The net take rate stood at approximately 4.4% in Q2 2024, a meaningful spread, but is under active pressure from the rate cap environment.

Pricing Power

Pricing power is essentially non-existent and structurally deteriorating. China's Supreme People's Court effectively imposed a hard 24% ceiling on borrowing costs across all private lending in October 2025, converting a judicial guideline into enforceable law. QFIN's historical ability to extract facilitation fees above this threshold — whether directly or through gray-zone bundled charges — has been explicitly targeted. The company cannot raise prices; it faces compulsory price reduction. This is the single most important fact about QFIN's current situation.

Geographic & Customer Concentration

Revenue is 100% China-derived from domestic operations. This is maximum geographic concentration in a single-country emerging market with a history of unpredictable regulatory interventions in fintech. No single borrower represents more than a de minimis share of revenue (the business is highly granular on the consumer side), but the institutional partner count of 167 financial institutions creates some exposure if key banking partners were to exit the platform. International expansion is nascent — management has begun "acquiring necessary credentials and setting up local organizations" in developing markets, but this represents no measurable revenue today.

Scale of the Business

Profitability

Cash Flow Quality

One genuine bright spot: operating cash flow of RMB 11.1 billion in FY2025 was described as a record, and Q4 alone generated RMB 3.15 billion. Cash from operations consistently exceeds GAAP net income, driven partly by timing differences in loan-related receivables and guarantee liabilities. FCF conversion is high because capital expenditures are minimal (technology platform business). Total cash and short-term investments stood at approximately CNY 10.72 billion (~US$1.5 billion) as of year-end 2025 — a substantial liquidity buffer relative to the ~$1.6B market cap.

Balance Sheet

The balance sheet has a dual character. On one hand, the platform's cash position is strong, provision coverage is robust at ~481–613% (different periods), and leverage at ~2.7x reflects the financial nature of the business. On the other hand, the capital-heavy credit business requires QFIN to either hold loans on its own books or provide guarantees to lending institutions — creating contingent liabilities that do not fully appear in simple leverage ratios. The company has issued convertible notes (2030 Notes) as part of a capital structure management exercise in March 2025. Off-balance-sheet guarantee obligations remain a meaningful risk: in a credit deterioration scenario, the guarantee release revenue that currently boosts profits could reverse into guarantee payment outflows. The 90-day delinquency rate of 2.71% in Q4 2025 and Day-1 delinquency of 6.1% warrant close monitoring as leading indicators of future credit losses.

Capital Intensity & ROIC

The business is fundamentally low-capex at the operating level — it is a software and data platform. Maintenance capex is de minimis. However, the capital-heavy credit-driven services segment effectively turns QFIN into a quasi-financial institution that must deploy capital as a co-investor in loans or as a guarantor, which is a structurally different capital requirement than pure software. Return on equity of ~24–30% in 2024 (varying by measurement method) is genuinely impressive and reflects the high-asset-turn nature of the facilitation model. ROIC is more difficult to compute cleanly given the blended nature of the business, but was above cost of capital throughout 2024. The trajectory into 2026 will depend entirely on whether regulatory pressure allows the economics to remain positive.

CEO Profile — Haisheng Wu

Haisheng Wu has served as CEO and Director of the company since its inception. His background is in digital products and user acquisition from his time at Baidu (BIDU), China's dominant search engine — a pedigree that speaks to digital distribution and consumer internet competency rather than traditional financial services expertise. He is a technology operator, not a financier. This matters in a business that is increasingly being forced by regulation to behave more like a bank. Wu participated in the UN's 80th anniversary events in New York in September 2025, suggesting continued international engagement at the highest level, and he is positioned as the public face of the company's international expansion thesis.

Founder Involvement

The company has origins connected to 360 Group (the Chinese internet security conglomerate), and a Seeking Alpha analysis from February 2026 noted "significant open-market purchases by QFIN's founder" — a high-conviction insider signal. The founder's willingness to buy shares in the open market near 52-week lows is one of the most bullish data points in this analysis. It suggests someone with the deepest information asymmetry believes the stock is cheap.

Board Composition

The board comprises Haisheng Wu, Alex Zuoli Xu, Dan Zhao as executive directors, plus Gang Xiao, Andrew Y. Yan, Fan Zhao, and Xiangge Liu as independent/non-executive directors. Andrew Yan is a known venture capital figure in China (SAIF Partners). The presence of directors linked to major institutional investors provides some governance legitimacy. The chairman is separate from the CEO. That said, as a Chinese ADR with a VIE structure, genuine board independence is structurally constrained — a universal caveat for this category of stock.

Skin in the Game

Insider buying has been unambiguously positive. A director-linked fund (TPC Compounder Master Fund) purchased 190,000 ADSs at $12.89 on March 27, 2026 — essentially at the 52-week low. The company repurchased ~$450M of its own stock in 2025 at an average price of $28.20 — significantly above current prices — which raises the question of whether buybacks were value-creative at those levels. The company has since indicated it will resume buybacks more actively when macro and regulatory conditions stabilize. Total capital returned to shareholders in 2025 was approximately $880 million (dividends + buybacks), representing roughly 98% of 2024 GAAP net income — a staggering capital return ratio that leaves virtually nothing for organic reinvestment.

Compensation Structure

Specific executive compensation details are available in the annual 20-F filing. Share-based compensation is excluded in non-GAAP measures, and the gap between GAAP and non-GAAP net income (~RMB 378M in FY2025) is attributable to SBC — moderate but not excessive relative to total income. The alignment between management pay and long-term shareholder value is difficult to fully assess without the proxy statement, but the pattern of aggressive buybacks suggests management is acutely aware of per-share value.

Management Credibility on Guidance

A significant credibility issue has emerged. When reporting Q4 2024 results in March 2025, management guided Q1 2025 non-GAAP net income of RMB 1.80–1.90 billion (49–58% year-on-year growth). This guidance was ambitious and suggestive of strong momentum. By Q4 2025, the company's operations had deteriorated sharply enough that Q1 2026 guidance was for RMB 900–950 million — a 51–53% year-on-year decline. The reversal from 58% guided growth to 53% actual decline within four quarters reflects either poor forward visibility or a regulatory environment that moved faster and harder than management anticipated. Either explanation is concerning for capital allocation confidence.

Does a Moat Exist?

A moat exists, but it is narrower than the company's market position and financial returns suggest, and it is eroding. The moat is real in the following senses: QFIN has proprietary AI-driven credit scoring infrastructure (the Argus Engine), seven-plus years of behavioral and repayment data on tens of millions of borrowers, established relationships with 167 financial institutional partners, and a consumer brand in Qifu Jietiao that commands 291 million registered users. These are genuine barriers. A new entrant would require years and billions of RMB to replicate them. The moat does not, however, protect against regulatory restructuring of the economic terms of the business.

Moat Type

Data/Intangible Asset Moat: The proprietary credit scoring data accumulated over seven years on Chinese thin-file borrowers is the core asset. Each new loan vintage improves the model's accuracy, creating a self-reinforcing data flywheel. This is the most defensible element of the moat.

Scale Economies: Operating leverage is real — 3,527 employees servicing a $2.75 billion revenue base implies roughly $780,000 revenue per employee, a remarkable ratio for a financial services entity. Fixed-cost infrastructure spread over a growing loan book creates cost advantages.

Limited Network Effects: There are weak two-sided network effects — more borrowers attract more lenders, and vice versa — but they are less powerful than true marketplace network effects (like payments or social networks) because loans are fungible commodities and financial institutions can replicate distribution through multiple competing platforms.

Moat Strength & Durability

On a 3–5 year horizon, the moat's durability is actively threatened by two forces: (1) State-backed financial institutions such as WeBank (Tencent's digital bank) and Ant Group's Huabei/Jiebei products have effectively unlimited capital, superior regulatory standing, and access to platforms (WeChat, Alipay) that dwarf Qifu Jietiao's distribution. They can and do replicate QFIN's use cases at scale. (2) The regulatory 24% rate cap structurally removes the economic buffer that allowed QFIN to profit from riskier borrower segments that banks wouldn't serve — compressing the business toward the segments where state-backed institutions can compete most aggressively.

Market Structure & TAM

China's consumer credit market is among the largest in the world by absolute volume. The country has hundreds of millions of potential credit consumers — estimates of the underserved thin-file population historically exceeded 500 million people. However, the addressable market for an independent private platform like QFIN has been structurally narrowed by regulators who now require platform lenders to partner with licensed banks and limit the total take-rate economics. The TAM in which QFIN can operate profitably has shrunk materially since 2020.

Regulatory Environment — The Central Risk

The regulatory history of Chinese fintech is the single most important factor for understanding QFIN. The sector went from roughly 6,000 P2P lenders in 2018 to approximately 29 by 2020 after a systemic regulatory purge. QFIN survived that purge by pivoting from a P2P model to institutional partnership. However, October 2025 saw a new escalation: the 24% interest rate ceiling was converted from a judicial guideline to a hard legal cap, and by early 2026, both QFIN and Yiren Digital were summoned by regulators over high customer complaint volumes. All lending platforms were simultaneously required to disclose all facilitation fees by August 1 — eliminating the gray-zone fee structures that had historically inflated platform revenue margins.

The regulatory arc is unambiguous: Chinese authorities intend to tighten private consumer lending, protect borrowers from high effective costs, and channel more consumer credit through state-controlled or state-adjacent institutions. This is not a temporary cycle; it reflects a fundamental policy orientation. QFIN is collateral damage in a campaign targeting genuinely predatory lenders, but the regulatory instrument (rate caps, fee disclosure) does not distinguish between predatory and responsible operators.

Competitive Landscape

Cyclicality

Consumer credit platforms are highly procyclical — loan defaults rise during economic downturns, impairing both the guarantee obligations QFIN carries and the appetite of institutional partners to originate loans through the platform. China's economic headwinds in 2024–2026 (property sector stress, weak consumer sentiment, deflationary pressures) directly translate into elevated default rates for QFIN's borrower base. The 90-day delinquency rate of 2.71% in Q4 2025 reflects this cyclical stress.

Is the Cheapness Real?

The headline multiples are genuinely, eye-wateringly cheap by any Western or emerging-market fintech benchmark. A 2.1× trailing P/E, ~1× forward cash flow, and 12% dividend yield on a company with $1.5 billion in cash generating $2.75 billion in annual revenue would normally be a screaming buy signal. The market is not being irrational here, however. It is pricing in a specific and credible set of risks.

The critical question is: what is the forward earnings power of this business after the 24% rate cap is fully implemented? The Q1 2026 guidance of RMB 900–950 million non-GAAP net income (annualized: ~RMB 3.6–3.8 billion) compared to FY2025's RMB 6.354 billion suggests the run-rate earning power has approximately halved. On a forward basis, the stock at $13 and ~$1.6B market cap (RMB ~11.2B) is trading at roughly 3× normalized forward earnings — which is still cheap if that trough is the floor, but is not obviously cheap if earnings continue deteriorating.

DCF Sanity Check

Using conservative assumptions: Revenue grows at 0% (flat from 2025 levels — optimistic given Q4 2025 volume trends); net margins compress from 31% to 20% reflecting full rate-cap impact (approximately RMB 3.84B net income on RMB 19.2B revenue); FCF conversion of 90%; discount rate of 15% (China-specific risk premium warranted); terminal growth of 2%.

Implied intrinsic value under these assumptions: approximately RMB 25–32B enterprise value, or roughly US$3.6–4.6B equity value after netting the cash position (~US$1.5B) and existing obligations. At ~US$1.6B market cap, this implies 2–3× upside in a "muddle through" scenario. At harsher assumptions (margins compress to 15%, revenue declines 10%), intrinsic value approaches current market cap. The margin of safety is material but not overwhelming given the uncertainty range.

Why the Stock Is Where It Is

The stock fell from ~$49 to ~$13 (-73%) for identifiable, fundamental reasons: (1) China's October 2025 hard 24% rate cap directly attacked QFIN's revenue model; (2) Q4 2025 profit fell nearly 50% year-on-year; (3) Q1 2026 guidance implies another 52% year-on-year decline; (4) Both QFIN and peers were summoned by regulators in early 2026 over customer complaints; (5) Macro deterioration in China raised credit losses. This is not a case of irrational sector-wide selling. The decline is driven by genuine fundamental deterioration interacting with a structural regulatory restructuring of the business's economics.

Value Trap Risk Assessment

The value trap risk is elevated but not conclusive. QFIN is not in structural decline from substitution or obsolescence — the underlying demand for consumer credit in China is enormous and will persist. The question is whether QFIN can survive in the regulatory environment with sufficient economics to remain a viable business. The cash position ($1.5B vs $1.6B market cap), low payout ratio (22%), and strong operating cash flow generation argue against a near-term existential threat. But if regulatory pressure continues to escalate, the business model may become permanently impaired in ways that permanently lower the earnings ceiling.

Dividends

QFIN has established a progressive semi-annual dividend policy and raised its dividend alongside the Q4 2025 results announcement — a signal of management confidence in the cash generation capability of the business. The H2 2025 dividend of US$0.78 per ADS (approximately $1.44–1.56 annualized) represents a yield of ~11–12% at current prices — a genuine income return that is not obviously at risk given the 22% payout ratio and strong cash generation. The dividend is well-covered and management has signaled intent to maintain it. This is one of the clearest bull-case elements: you are being paid 12% to wait.

Share Buybacks — A Mixed Record

The buyback record is substantial but raises execution questions. Between June 2023 and December 2025, QFIN repurchased approximately 40 million ADSs — roughly 25.4% of the starting share count — at an aggregate cost of approximately US$950–1,000 million (2023 plan at avg $16.02, 2024 plan at avg $23.14, 2025 plan at avg $28.20). The 2025 buybacks at an average of $28.20 are now 54% underwater against the current ~$13 price. The capital destruction from poorly-timed buybacks is significant: QFIN spent roughly $450 million in 2025 buying stock at $28.20 that is now worth approximately $207 million — a paper loss of ~$243 million. That said, from a per-share value standpoint, the share count reduction has meaningfully boosted EPADS even as absolute earnings declined. Management has said it will resume buybacks more actively once conditions stabilize — at $13, the value proposition is considerably more compelling than at $28.

M&A Track Record

QFIN has not been an aggressive acquirer. The capital has been directed primarily toward the buyback program and dividends. The international expansion being pursued in 2025–2026 (acquiring credentials in developing markets) is organic, not M&A-driven. This conservative stance on acquisitions is a positive governance signal — the company has not destroyed value through empire-building.

Stated Strategic Priorities

Management has articulated three primary strategic responses to the regulatory and macro headwinds: (1) Tighten risk standards — raise the credit quality bar for new borrowers, accept lower volumes in exchange for better loss rates. Early data is encouraging: December vintage FPD30 was near 2-year lows, and January 2026 FPD7 improved ~10% from December. (2) Accelerate the capital-light business mix — pivot away from the capital-heavy credit-driven service model toward the ICE and technology solutions models where QFIN bears no credit risk and earns pure facilitation fees. Ironically, Q4 2025 showed ICE balances declining 25.3% YoY, suggesting this pivot is not yet executing as planned. (3) International expansion — establish operations in developing markets to diversify away from China regulatory concentration.

Early Evidence of Progress

The tightened risk standards appear to be working at the new-vintage level. However, the legacy portfolio continues to deteriorate — C-M2 delinquency of 0.97% in Q4 and the 90-day delinquency of 2.71% reflect stress on existing loans originated under looser prior standards. The improvement in new vintages will take 12–18 months to flow through to reported financials. International expansion is too early-stage to be financially material. The capital-light pivot is the most strategically important initiative and the most uncertain in execution.

Potential Catalysts (12–24 Months)

The most significant positive catalyst would be regulatory clarity: if the NFRA and Supreme Court guidance is formalized in a way that creates a stable operating framework — even a lower-revenue one — QFIN's earnings would stabilize and the stock would likely re-rate sharply. A secondary catalyst is international revenue contribution becoming material. A third catalyst is the Q1 2026 earnings release (expected May 27, 2026): if actual results exceed the guided RMB 900–950M range, it would signal that the trough has passed. A negative catalyst (further regulatory escalation, credit loss acceleration, ABS market disruption) would likely push the stock to new lows.

AI as the Core Business

Unlike most companies in this report framework where AI is an adjacency or a threat, AI is literally QFIN's product. The Argus Engine, the intelligence credit engine (ICE), and proprietary risk management models are the reason QFIN can underwrite borrowers that traditional banks cannot profitably serve. Every loan facilitated is, in a sense, an AI inference. A co-authored academic paper was accepted at IJCAI 2025 (a top-tier AI conference), and the company was granted "Most Honored Company" recognition in fintech AI circles in May 2025 — signs of genuine R&D engagement rather than PR-driven AI-washing.

AI as a Revenue Opportunity

The "total technology solutions" segment — end-to-end technology licensing to financial institutions (on-premise deployment, SaaS, hybrid) launched in 2023 — represents QFIN's bid to monetize its AI models as a software product independent of loan origination. If this segment scales, it would provide genuinely recurring, non-credit-risk revenue that is far less exposed to interest rate caps and delinquency cycles. Currently, this segment is small and grouped within Platform Services. It is the most strategically interesting long-term growth vector, but it remains unproven at meaningful revenue scale.

AI as a Threat

The threat from AI to QFIN is that large language models and increasingly capable foundation models could allow state-backed banks to rapidly close the AI underwriting gap that currently gives QFIN its advantage. WeBank (Tencent) and Ant Group have deep AI capabilities and superior data assets (from WeChat and Alipay behavioral data). The competitive moat that QFIN holds in AI credit scoring may have a shorter remaining lifespan than the 2018–2024 era suggested. QFIN's seven years of credit behavioral data is valuable, but Tencent's WeChat behavioral data on a billion users is arguably more valuable as an AI training asset.

Technology Investment Posture

With only 3,527 employees and $2.75B in revenue, QFIN operates with extreme capital efficiency. R&D spending is not separately broken out at a headline level, but the company's asset-light structure implies a meaningful proportion of the workforce is technology-focused. The ongoing work with Beijing Jiaotong University (joint AI research milestone announced May 2025) and IJCAI paper acceptance suggest a genuine commitment to advancing the technical frontier rather than just maintaining existing systems.

Insider Activity

The insider signal is unambiguously bullish in recent weeks. Director Xiaohuan Chen's affiliated fund (TPC Compounder Master Fund) purchased 190,000 ADSs at $12.89 on March 27, 2026 — near the 52-week low — while simultaneously closing out short put positions (suggesting they are removing hedges and increasing net long exposure). This is a high-quality insider signal: not routine grant-based ownership, but discretionary open-market purchase at a price that represents maximum pessimism. Earlier reports also note significant open-market purchases by QFIN's founder near recent lows.

Institutional Activity

The most striking institutional data point is Oversea-Chinese Banking Corp (OCBC) — Singapore's conservative, highly regulated state-linked bank — increasing its QFIN stake by approximately 742%, as noted in a February 2026 Seeking Alpha analysis. OCBC is not a speculative hedge fund; a near-8× position increase by a conservative Asian institutional investor represents genuine value-oriented conviction. This type of smart money accumulation at current prices is a meaningful signal.

Short Interest & Analyst Consensus

Short interest is low at approximately 1.62% of outstanding shares, down 30% in April 2026. The low short interest is somewhat surprising given the stock's decline — it suggests short sellers are not aggressively pressing the bear case at current levels, possibly because the valuation has already priced in considerable bad news. Analyst consensus is mixed: Jefferies maintains a Buy with a $23.40 target (implying ~80% upside); average price target across analysts is ~$22.20 (implying ~70% upside from $13); one source notes a consensus "Reduce" rating. The divergence between analysts — some with high price targets and some with Reduce ratings — reflects genuine fundamental uncertainty, not analyst complacency.

Asymmetry Assessment

The base case offers approximately 50% upside and the bear case implies 45% downside — roughly 1.1:1 risk/reward, which is below the threshold for a high-conviction investment. The bull case of 2–3× upside moves the asymmetry to a more attractive range, but requires a specific and uncertain catalyst (regulatory clarity). The uncertainty around the Q1 2026 earnings release (May 27, 2026) and the broader regulatory arc makes the near-term risk/reward unattractive for investors who require clarity before committing. For investors with high China political risk tolerance and a true 3–5 year time horizon, the math looks different — the business is unlikely to be worth zero, and the cash alone provides significant downside protection.