Trip.com Group Limited (NASDAQ: TCOM; HKEX: 9961) — A 5.7% Toll Booth on Trial for Monopoly, and a 7× P/E That Isn’t One

Report date: 25 July 2026 · Reference price: US$43.64 (NASDAQ close, 24 July 2026)

Coverage initiation. Trip.com Group is a Cayman Islands holding company and a foreign private issuer: it reports on Form 20-F and Form 6-K, not 10-K/10-Q, and its financial statements are presented in Renminbi. All USD conversions in this memo use RMB 6.8905 = US$1.00, the company’s own convenience translation rate in its Q1 2026 release, unless otherwise stated.

⚡ Claude’s Take

This block is the author’s own independent opinion, offered as general information and not investment advice. It is the only place in this article where a position is taken: the analysis in sections 1–15 below carries no recommendation and no price target of any kind.

Verdict: HOLD at US$43.64 — a defensible accumulation zone of roughly US$36–44, which means the stock is already at the upper edge of where I would be buying rather than comfortably inside it. Estimated fair-value zone US$53–60 on base-case normalised earnings (≈12× EV/EBIT, the low end of Trip.com’s own eleven-year multiple range). Not a short at any price I can defend, and not a table-pounding buy either. Conviction: medium.

The screen says Trip.com trades at 7.0× earnings, the 0.9th percentile of its own ten-year history. That number is wrong, and so is management’s own non-GAAP version of it. In June 2025 Trip.com sold 34.4 million MakeMyTrip shares back to MakeMyTrip for ~US$3.0 billion and booked a RMB 15.2 billion gain. The company’s non-GAAP bridge strips out fair-value changes on equity securities — but a realised disposal gain is not a fair-value change, so it flows straight through to “non-GAAP” EPS. Management gave the honest figure verbally on the Q4 call and nowhere else: “Excluding gains from investments, net income attributable to Trip.com Group Limited was RMB 13.4 billion” — against RMB 33.3 billion reported. That is US$2.75 per ADS, or ~15.9× — not 7×. Anyone underwriting this on the screen multiple is underwriting a one-time asset sale as though it were the business.

What survives that correction is still interesting, and it rests on one number the sell side does not foreground: Trip.com’s blended take rate is 5.7% (RMB 62.4bn of revenue on RMB 1.1 trillion of gross bookings) — against Booking’s 14.5%, Airbnb’s 13.4% and Expedia’s 12.3%, each computed from those companies’ own FY2025 disclosed gross bookings and revenue. Trip.com is the least extractive major online travel agency in the world, by a factor of more than two, and it is the one being investigated for abusing a dominant position. The conduct at issue is real and indefensible — an AI tool that scanned rivals’ prices and forced hotel partners to cut their own, which Trip.com shut down on 10 March 2026 — but it is a conduct problem, not a rent-extraction problem, and the distinction matters enormously for what the remedy can cost. On the Alibaba (4% of China revenue) and Meituan (~3%) precedents, a fine lands around RMB 1.1–1.5 billion (US$160–215m) — roughly 2% of market capitalisation and about six weeks of gross cash. The fine is not the issue. The permanent loss of the price-adjustment tool and of rail value-added-service monetisation is the issue, and at my sizing it costs 15–30% of operating income. At US$16.9 billion of enterprise value against US$2.35 billion of trailing operating income — 7.2× EBIT, or 5.3× stripping the US$4.6 billion strategic-stake portfolio at book — the market is discounting a 40% permanent decline in operating profit. That is between 1.3× and 3.7× the damage I can actually construct from the disclosure, from a company whose Q1 revenue still grew 17%, whose international platform is compounding at 60–65%, and which holds US$10.5 billion of net cash equal to 38% of its market capitalisation.

The framing is deep value / falling knife, and I want to be honest that it is the second of those as much as the first. This is not a quality compounder on temporary sale. Every risk-adjusted horizon is negative — 3-month –54%, 6-month –52%, 1-year –32% annualised, with a 1-year Sharpe of –0.90 and a ten-year annualised return of +0.17%. The factor model is even more damning about what this security actually is: its largest loadings are Country: China (+0.94) and Industry: Social Media (+1.09), its factor-similar peers are Xiaomi, Meituan, Baidu, JD and PDD with not one Western travel name in the list, and its loading on the “Travel Leisure Giants” basket is +0.010 — indistinguishable from zero. The tape does not price this as a travel company. It prices it as a China-internet ADR, and that is precisely why the de-rating is idiosyncratic (10.6th own-history percentile against 35th–64th for every OTA peer, same date) rather than a travel-sector call. Two other facts cut in opposite directions and both belong here: management repurchased roughly US$1.0–1.2 billion of stock in Q1 2026 alone, nearly double the entire 2025 programme, buying into its own regulatory shock — the best-informed buyer voting with the balance sheet. And in the first week of September 2025, within days of the +14.9% earnings pop and within 10% of the all-time high, insiders — including 1,000,000 ADS of founder-family “Founder Shares” — sold ~US$106.8 million. Both are real. Neither cancels the other.

Conviction: medium. The single fact that flips me bullish: the Q2 print (late August 2026) showing revenue at the top of the 3–8% guide and Q3 guidance back above 10% — proving the reset was a level rebasing rather than a rate change. The single fact that flips me bearish: Q3 guided in the low single digits as well, or an international-platform growth rate falling below 40% — either would mean the compliance reset is structural and the China domestic take rate is being permanently re-based, at which point 7× EBIT is not cheap, it is correct. Tag: “The lowest toll in travel, on trial for charging too much.”

📈 Stock Price Action — Five-Year Event Map

Trip.com has completed a full five-year round trip and is now in the drawdown phase of it. The ADS bottomed at US$16.67 on 14 March 2022 in the China-ADR delisting panic, compounded almost fivefold to an all-time high of US$78.96 on 12 January 2026 on the post-COVID travel recovery and the international-expansion story, and has since fallen –44.7% to US$43.64. The 52-week range is US$39.84–78.96; the low was set on 30 June 2026, three weeks ago. The stock trades below its 21-, 50- and 200-day exponential moving averages (US$43.17 / US$45.38 / US$54.28). The entire decline is the product of two dated, identifiable, company-specific events — not a travel-demand deterioration.

Cycle narrative. (1)–(2) The 2021–22 collapse was not about Trip.com; it was the China-ADR complex de-rating on the platform crackdown and on genuine fear that US-listed Chinese shares would be forcibly delisted under the Holding Foreign Companies Accountable Act. Trip.com’s own revenue was flat at RMB 20.0 billion across both 2021 and 2022 under zero-COVID, and it lost money in 2021. (3) The single largest up-day in five years came from a government statement, not from the company — an early, clean illustration of how much of this security’s variance is policy, not operations. (4)–(5) The Party Congress and zero-COVID drove one more leg down before the abrupt December 2022 reopening nearly doubled the stock in two months; revenue then went RMB 20.0bn → 44.5bn (2023) → 53.3bn (2024) → 62.4bn (2025). (6) The September 2024 stimulus package re-rated the whole China complex. (7) The August 2025 beat took the stock to a +14.9% day and on to its all-time high — and it is worth noting that in the first week of September 2025, days after that pop, insiders sold ~US$106.8 million including 1.0 million ADS of founder-family shares (see Capital Allocation). (8) On 14 January 2026 the company disclosed a SAMR investigation under the PRC Anti-Monopoly Law into whether it “has abused or is abusing a dominant market position”; the ADS fell 17.0%, its second-largest down-day in five years, and the underlying conduct — an AI tool forcing hotel partners’ prices down — had surfaced publicly in late November 2025. (9) On 24 June 2026 Trip.com reported Q1 revenue +17% but guided Q2 to +3–8% with margin compression, citing macro headwinds and “operational adjustments… to align with evolving industry standards and compliance frameworks”; the ADS fell 12.5% on the close and as much as 18% intraday, reaching a 52-week low six days later. (10) The stock has recovered ~9.5% off that low on no new information.

Price moves are Fact; the attributed drivers are Interpretation. No recommendation, price target, or technical level is expressed or implied in this section.

1. Executive Summary

Trip.com Group is the dominant online travel agency in China and, increasingly, in Asia-Pacific. In FY2025 it intermediated approximately RMB 1.1 trillion (US$160bn) of gross bookings across four brands — Ctrip (China full-service), Qunar (China value), Trip.com (international) and Skyscanner (metasearch) — converting that into RMB 62.4 billion (US$9.1bn) of net revenue, +17% year-over-year, at a blended take rate of 5.7%. Operating income was RMB 15.8 billion (25.3% margin), gross margin was 80.6%, capital expenditure was 1.3% of revenue, and the balance sheet carries US$10.5 billion of net cash — 38% of market capitalisation. The business is asset-light, structurally profitable, funded by negative working capital (a –446-day cash conversion cycle), and growing faster than any Western OTA peer: international platform gross bookings rose ~60% in FY2025 and ~65% year-over-year in Q1 2026, with inbound-to-China bookings up ~90%.

Three facts dominate the current situation. First, reported earnings are badly distorted. FY2025 net income of RMB 33.3 billion contained RMB 21.3 billion of “other income,” of which RMB 15.4 billion was the gain on selling a MakeMyTrip stake back to MakeMyTrip for US$3.0 billion. Management’s own non-GAAP measure does not remove this — it excludes fair-value changes, and a realised disposal is not a fair-value change — so both the GAAP P/E (7.0×) and the non-GAAP P/E (6.7×) are wrong by roughly 2.4×. Management’s own clean number, given only verbally on the Q4 call, is RMB 13.4 billion, or ~15.9× at the reference price.

Second, the company is under a live PRC Anti-Monopoly Law investigation with no reserve recorded. SAMR opened it in January 2026 over an AI price-adjustment tool that scanned competitors’ rates and forced hotel partners to cut their own prices — economically the same price-parity enforcement the EU banned for Booking.com under the Digital Markets Act. Trip.com shut the tool down on 10 March 2026. A separate June 2026 joint action by SAMR, the Cyberspace Administration and the National Railway Administration targeted rail-ticket “acceleration package” monetisation across seven platforms. The company cannot estimate a loss and has not reserved one; the 20-F states the outcome “may have a material adverse effect.”

Third, the cost is already visible in guidance. Q2 2026 revenue is guided to +3–8% from +17% in Q1, explicitly attributed in part to “operational adjustments… to align with evolving industry standards and compliance frameworks.” Our sizing puts the permanent compliance-related cost at 15–30% of operating income. The market, at 7.2× EV/EBIT (5.3× stripping the strategic-stake portfolio), is discounting roughly 40%.

The competitive position is genuine but narrower than the market share implies. Trip.com’s moat is a supply-side scale and fulfilment advantage — decades of direct integrations with Chinese hotels, airlines and the rail system, 16,718 customer-service employees delivering 24/7 support in 30+ languages, and a one-stop product surface competitors cannot replicate cheaply. Under Greenwald’s taxonomy this is economies of scale combined with modest customer captivity, not network effects. But Chinese consumers multi-home aggressively across Meituan, Douyin, Fliggy and Tongcheng, and the 5.7% blended take rate is itself evidence that the company has never been able to price like a monopolist. The uncomfortable corollary: a platform earning a 5.7% toll has less cushion to give back than one earning 14.5%.

Capital allocation has improved markedly and is the strongest management evidence in the file: a US$5 billion repurchase authorisation in August 2025, a first-ever dividend in 2025, and an inferred US$1.0–1.2 billion of Q1 2026 buybacks executed into the regulatory selloff. Against that sit an option-heavy incentive structure (the executive chairman’s genuine equity stake is ~1.5% of shares against ~34 million options), a 5.0% diluted-share overhang, a 29% decline in free cash flow during a 17% revenue-growth year, a consolidated ROIC of 7.4% weighed down by RMB 75 billion of legacy goodwill and intangibles, and insider sales of ~US$106.8 million within 10% of the all-time high.

This memo takes no position and sets no price target — those are confined to Claude’s Take above. What follows is the mechanism, the numbers, and the falsification tests for each side.

2. Business Overview

2.1 What the company is

Trip.com Group Limited was founded in Shanghai in 1999 as Ctrip.com, listed on NASDAQ in 2003, renamed Trip.com Group in October 2019, and completed a secondary listing on the Hong Kong Stock Exchange (9961) in 2021. Its registered office is in the Cayman Islands and its principal executive office is at 30 Raffles Place, Singapore. It employed 43,574 people at 31 December 2025 (41,073 at YE2024; 36,249 at YE2023), a substantial majority in China.

It is a pure intermediary. It owns no hotels, no aircraft and no trains. It aggregates fragmented travel supply, aggregates fragmented traveller demand, and takes a toll on the transaction — the same structural position as Booking, Expedia and Airbnb.

2.2 The four brands

Qunar (consolidated December 2015) and Skyscanner (December 2016) are the source of the RMB 62.2 billion of goodwill on the balance sheet. Qunar Cayman Islands Limited is notable: 57% of it is owned by outside investment entities (M Strat Holdings, Momentum Strategic Holdings, Ocean Management, Earthly Paradise Investment Fund) yet it is consolidated in full under US GAAP because Trip.com is the primary beneficiary.

Over 90% of total transaction orders were executed through mobile channels in FY2025.

2.3 How it makes money — and the take-rate decomposition that matters

The headline operating metric is gross bookings; revenue is the company’s cut. For FY2025 management disclosed the components on the Q4 call, and they are worth setting out precisely because the blend conceals two very different businesses:

Three observations follow, and each does real work later in this memo.

(i) Air is half the gross bookings and a quarter of the revenue. Chinese domestic airfares are heavily administered and airline commissions were cut to near-zero in the 2015–16 “tiqian” reforms; a 2.6% take on RMB 550 billion is a volume business, not a profit pool. It exists to acquire and retain the customer.

(ii) Non-air transportation is RMB 8.2 billion — 13.1% of group revenue — and it is the line under direct regulatory attack. Reported “transportation ticketing” revenue was RMB 22.5 billion; management separately disclosed air-only at RMB 14.3 billion. The RMB 8.2 billion difference is rail, bus, ferry, ground transport and their associated value-added services. Chinese rail tickets are sold at government-fixed prices through the state 12306 platform with zero commission — so essentially all of the rail economics sit in the value-added services (priority “ticket-grabbing,” “acceleration packages,” bundled insurance) that regulators summoned seven platforms about in June 2026. Rail is a subset of that RMB 8.2 billion and cannot be isolated from public disclosure, but 13.1% of revenue bounds the exposure. (FACT: the RMB 8.2bn figure and its derivation. INTERPRETATION: that rail VAS is the economically meaningful part of it. OPEN QUESTION: the rail-specific split.)

(iii) The blended 5.7% take rate is the lowest of any major global OTA, by more than 2×. Against the peer set, each computed from FY2025 disclosed gross bookings and revenue: Booking 14.5%, Airbnb 13.4%, Expedia 12.3%. Even Trip.com’s accommodation-only take rate of 9.3% sits ~5 points below Booking’s blended rate. This is the single most important structural fact about the company and is developed in the Competitive Position and Variant Perception sections below.

2.4 Recurring vs. non-recurring

There is no subscription revenue. Every yuan is transactional, recognised at or near the time of travel. What substitutes for contractual recurrence is habit and breadth: a user who books rail on the platform monthly and hotels quarterly is functionally recurring, and management’s entire strategic argument for the low-take transportation business is that it is the acquisition and frequency engine for the high-take accommodation business. That argument is credible — but it is also exactly what makes the rail crackdown more strategically important than its 13.1% revenue share suggests, because it degrades the top of the funnel, not just a revenue line.

2.5 Geographic mix — the genuine growth story

International contributed ~40% of revenue and bookings in FY2025, up from ~35% in FY2024. Within that:

• International OTA platform (Trip.com brand) gross bookings: +60% in FY2025, +65% year-over-year in Q1 2026
• Inbound-to-China bookings: +90% year-over-year in Q1 2026; ~20 million inbound travellers served in 2025, ~7 million in Q1 2026 alone
• Management target: 200 million inbound travellers over five years

Management frames inbound as the structural opportunity: inbound tourism is ~0.5% of China’s GDP against >10% for Thailand and 5–6% for France, Italy and Spain, implying “at least 5 to 10× growth.” The policy backdrop is genuinely supportive — 80+ countries with visa-free access, 10-day visa-free transit, and Alipay/WeChat Pay now accepting foreign cards. This is the one part of the business where the state is a tailwind rather than a headwind, and it deserves weight.

Verdict. A clean, asset-light, high-gross-margin transaction-toll business with a dominant domestic franchise, a genuinely fast-growing international arm, and an unusually low take rate. The model is sound and simple to understand. The structural vulnerability is that the low-take transportation business — half of gross bookings, a third of revenue — is administered-price inventory whose only real economics sit in value-added services that regulators have now targeted.

3. Industry Dynamics

3.1 Market structure

China’s travel market is the largest single-country travel market by volume and among the most online-penetrated. Its structure differs from the West in four ways that matter for economics:

(a) Administered pricing on transport. Rail is a state monopoly (China State Railway Group) selling through 12306 at fixed prices with no commission and no acceleration fees. Domestic air is dominated by three state-controlled carriers with heavily constrained commissions. Roughly half of Trip.com’s gross bookings therefore sit on inventory it cannot meaningfully monetise.

(b) Lower hotel commissions. Chinese hotel commissions run structurally below Western norms — Trip.com’s 9.3% accommodation take rate against Booking’s 14.5% blended. Chinese hotel supply is more fragmented and more independent (which should support higher commissions), but it is also more contested, with Meituan able to cross-subsidise hotel distribution from its food-delivery and local-services flywheel.

© Super-app competition. The most dangerous competitors are not travel companies. Meituan attacks from local services with enormous low-tier-city hotel share; Douyin (ByteDance) attacks from short-video and livestream commerce, selling hotel and destination vouchers at the point of inspiration; Fliggy (Alibaba) attacks from e-commerce. Each has a demand-acquisition asset Trip.com must rent.

(d) The regulator is an active participant, not a referee. This is the defining structural feature and is treated immediately below.

3.2 Market size and profit pools

The industry’s profit pool sits overwhelmingly in accommodation. On Trip.com’s own numbers, accommodation is 25% of gross bookings and 42% of revenue; air is 50% of bookings and 23% of revenue. Applying Trip.com’s group operating margin uniformly would understate accommodation’s contribution and overstate transportation’s materially — transportation at a 2.6% take on inventory requiring the same fulfilment infrastructure is, at best, marginally profitable and plausibly loss-making before the value-added services now under threat.

Growth drivers are real and multi-year: rising per-capita discretionary income; outbound travel still below its 2019 peak in some corridors; inbound travel from a very low base with explicit policy support; and the “silver generation” (Old Friends Club hotel bookings +100% year-over-year in Q1 2026) and entertainment-driven travel (+74% in Q1 2026) as genuinely incremental occasion categories.

3.3 Regulation — the dominant variable

Three distinct regulatory threads are live, and conflating them produces the wrong answer.

Thread 1 — the SAMR Anti-Monopoly Law investigation (opened January 2026). SAMR is investigating “whether the Company has abused or is abusing a dominant market position to engage in monopolistic conduct.” The conduct became public in late November 2025: hotel partners reported that Trip.com’s automated AI price-adjustment tool scanned competitors’ prices and forced reductions on the partners’ own listings, with non-compliance punished by reduced search visibility or delisting. Partners characterised it as “one-sided coercion” and a loss of pricing autonomy. Trip.com shut the tool down on 10 March 2026.

The correct analytical frame is Booking.com: this is algorithmic enforcement of price parity. Booking’s contractual parity clauses were banned in the EU under the Digital Markets Act — a genuine erosion of Booking’s supply-side moat. Trip.com achieved the same outcome by software rather than contract, and it is being dismantled by a different regulator on different legal grounds. The economic consequence is identical: the platform loses its guarantee of holding the lowest price on the supplier’s inventory, which weakens the “book here, it’s cheapest” proposition and reduces conversion.

Thread 2 — rail ticketing (June 2026). On 12 June 2026, SAMR, the Cyberspace Administration of China and the National Railway Administration jointly summoned seven online travel platforms — including Trip.com, Qunar, Meituan, Tongcheng and Fliggy — demanding rectification of deceptive train-ticket sales practices: marketing “acceleration”/“ticket-grabbing” packages that cannot in fact bypass the official queue, steering passengers to mismatched routes, and improper personal-data collection. The state 12306 channel carries no markup, no acceleration fee, no bundled insurance and no membership fee. CFO Cindy Wang confirmed the exposure on the Q1 call: “optimization of certain rail-related products and value-added services may create some near-term headwinds… already reflected in our current expectations and partially incorporated into our Q2 outlook.”

Thread 3 — the standing VIE and ADR overhang. Trip.com is a Cayman holding company operating in China partly through variable interest entities. Worth noting as a genuine mitigant: the VIEs constituted only 5% of total assets at 31 December 2025 — far below the proportion at Alibaba, Baidu or JD. The three significant VIEs (Ctrip Commerce, Shanghai Huacheng, Qunar Beijing) hold the value-added telecommunications and travel-agency licences; the great majority of the economics sit in directly-owned PRC subsidiaries. The Hong Kong dual listing further reduces forced-delisting risk relative to single-listed ADRs.

3.4 Where the industry sits in the capital cycle

Applying Marathon’s supply-side lens: the Chinese OTA industry is not in a capital-attraction phase. Capital intensity is trivial, but the binding constraint on new entry is demand acquisition and supply integration, both of which are expensive and slow. There has been no meaningful new scaled entrant in a decade — the competitive pressure comes from adjacent incumbents (Meituan, Douyin, Alibaba) redeploying existing demand assets, which is a different and more dangerous phenomenon than new capital entering. The regulator is now doing something the capital cycle would not: forcibly reducing the returns of the incumbent to redistribute them to suppliers and consumers. That is a policy-driven return compression, and it will not mean-revert on its own.

Verdict: a structurally attractive industry made structurally less attractive by an active, interventionist regulator. The demand growth is real and multi-year; the intermediary position is genuinely valuable; capital intensity is negligible and returns on tangible capital are extraordinary. But roughly half the gross bookings sit on administered-price inventory that cannot be monetised, hotel commissions are structurally below Western levels, the demand-acquisition layer is contested by three better-capitalised super-apps, and the state has demonstrated both the willingness and the mechanism to reset platform economics by decree. A good industry, in a jurisdiction that caps how good it is allowed to be.

4. Competitive Position

4.1 Naming the moat

Under Greenwald’s taxonomy, Trip.com’s advantage is economies of scale combined with modest customer captivity — not network effects, and not a cost advantage in the classic sense.

The scale mechanism is real and measurable. Trip.com carries fixed costs that are large in absolute terms and trivial per transaction: 18,029 product-development staff, 16,718 customer-service staff delivering 24/7 support in 30+ languages, direct technical integrations with hundreds of thousands of suppliers, and a fulfilment infrastructure that must handle rebooking, cancellation, visa support and crisis response across 206 countries. A challenger must replicate all of it before it can compete on the full product, and it must do so to serve a customer whose lifetime value is capped by a 5.7% take rate. That is a genuinely unattractive entry proposition, and it explains the absence of new scaled entrants.

The captivity mechanism is weak, and management’s own evidence shows it. Asked directly about competition on the Q4 call, CEO Jane Sun’s answer was entirely about service quality: 24/7 call centres, comprehensive product, global coverage, evacuating customers from conflict zones within two minutes. These are real differentiators, but they are operational excellence, not structural captivity — a competitor can hire call-centre staff. There is no switching cost of consequence: Chinese travellers multi-home aggressively, price-compare across Meituan, Qunar, Fliggy, Tongcheng and Douyin, and the loyalty programme is unremarkable.

4.2 The Greenwald market-share-stability test

Greenwald’s test is that stable share over five to eight years implies a barrier to entry, and share shifts greater than ~5 percentage points imply there is none. Trip.com is estimated to hold roughly 48% of China’s online hotel booking market, a share that has been broadly stable for several years despite sustained assault by Meituan and Douyin. On the domestic hotel franchise, the test is passed — this is the strongest single piece of moat evidence available.

But that stability must be read alongside a second fact: the 5.7% blended take rate. A firm with a genuine, unconstrained barrier to entry converts it into price. Trip.com holds roughly half its core market and earns less than 40% of Booking’s toll. Either (a) the barrier is real but the industry structure caps what it can extract — administered transport pricing, fragmented low-ADR hotel supply, super-app cross-subsidy — or (b) the barrier is thinner than the share implies and the price is competed away. The evidence supports (a) as the dominant explanation, with a real contribution from (b) in the contested low-tier-city segment where Meituan is strongest.

The irony worth stating plainly: Trip.com is being investigated for monopolistic abuse while earning less than half the toll of the least-regulated Western OTA. That is not a legal defence — SAMR’s case is about conduct (coercing suppliers’ pricing autonomy), not about the level of the take rate, and coercion is coercion at any price. But it is a powerful economic observation, and it bounds how much a remedy can plausibly cost: you cannot reset a take rate that was never elevated.

4.3 Head-to-head

Peer figures are from those companies’ FY2025 reported results and are not struck at the same date as Trip.com’s market data; they are used for structural comparison, not for a live relative-value calculation.

The table says something precise. Trip.com grows fastest, is the least extractive, holds by far the strongest balance sheet, and earns the lowest consolidated return on capital. The first three are franchise facts. The fourth is a capital-allocation fact — RMB 75.2 billion of Qunar/Skyscanner goodwill and intangibles plus a RMB 31.8 billion equity portfolio sitting on top of an operating business that needs almost no capital at all (see Financial Quality).

4.4 The AI question

Both management and the market treat agentic AI as an existential question for OTAs. Executive Chairman James Liang’s answer is the most coherent offered by any OTA management: the OTA model rests on inspiration, transaction and service; AI agents excel at inspiration, which is the least defensible layer, and their rise “reinforces the critical importance of our transactional and service layers.” Trip.com’s response is two-sided — building its own vertical travel models and agentic search, while simultaneously opening the platform to third-party agents via Skills, MCP interfaces and agent-to-agent transaction rails, aiming to be “the trusted infrastructure for AI agents.”

The honest assessment: this is the correct strategy and it is also an admission. If AI agents become the demand-acquisition layer, Trip.com’s economics migrate from owning the customer to supplying the inventory — precisely the margin leakage the OTAs already suffer to Google in the West. Trip.com is somewhat better insulated than Booking or Expedia because Chinese demand acquisition runs through super-apps rather than Google search, and because its 5.7% take rate is a much smaller prize for an agent to disintermediate. But it is not immune, and a 2.6%-take air business is exactly the kind of commoditised fulfilment an agent would happily route around.

Verdict: a real, financially-visible competitive advantage — scale in supply integration and fulfilment, validated by a stable ~48% share of Chinese online hotel bookings against relentless super-app attack — that is nonetheless structurally capped rather than merely contested. Applying the test that a moat must tie to a financial outcome that would deteriorate without it: remove Trip.com’s supply scale and service infrastructure and the accommodation take rate converges toward the contested rate Meituan can offer, and the fulfilment guarantee that justifies the toll disappears. That is a genuine, quantifiable degradation, so the moat is real. But a moat that yields a 5.7% blended toll while a lesser-moated Western peer yields 14.5% is a moat whose owner never controlled the price. And the regulator has now demonstrated it can narrow it further by decree.

5. Growth History and Forward Opportunities

5.1 The record

*Company-basis operating income (excludes “other operating income” that some data providers fold in). See the Financial Quality section.

Two things stand out. The COVID trough was brutal but the franchise survived intact — gross margin never fell below 77% even at RMB 18.3 billion of revenue, evidence of a genuinely variable cost structure. And revenue in 2025 was only 75% above 2019 after six years, a compound rate of ~9.8% across the full cycle. The post-COVID growth rates (+122%, +19.7%, +17.1%) are recovery arithmetic decaying toward a structural rate, and the honest read of the trend is that the structural rate was converging toward low-to-mid teens before the compliance reset.

5.2 Decomposition of FY2025 growth

Growth is entirely organic — there has been no material acquisition since Skyscanner in 2016. Mix is improving: the high-take accommodation line grew ~2× the low-take transportation line, which is exactly the mix shift a shareholder wants. Q1 2026 continued it (accommodation +17%, transportation +12%, packaged tour +19%, corporate +20%).

5.3 Forward opportunities, ranked by credibility

(1) Inbound travel to China — highest credibility, largest optionality. Inbound bookings grew ~90% year-over-year in Q1 2026 off ~20 million travellers in 2025. The policy tailwind is concrete and verifiable: 80+ visa-free countries, 10-day visa-free transit, foreign-card acceptance on Alipay/WeChat Pay. The 0.5%-of-GDP starting point against Thailand’s 10%+ is a genuine gap. Trip.com is uniquely positioned because it owns both the Chinese supply relationships and, through Trip.com and Skyscanner, the international demand channel — a combination neither Booking nor Meituan possesses. Management invested >RMB 1 billion in inbound in 2025 and connected 63,000 hotels to their first-ever inbound booking. This is the single most differentiated growth asset in the company and it is the part the market is currently paying nothing for.

(2) International OTA (Trip.com brand) in APAC — high credibility, contested. Gross bookings +60% (FY2025) and +65% (Q1 2026), international now ~40% of revenue against ~35% a year earlier. Management states the international segment’s margin profile has “significantly improved” and expects a “structurally stronger profitability profile.” The contest is direct with Booking’s Agoda, which is entrenched across Southeast Asia — a UBS/HSBC analyst asked about exactly this on the Q4 call and management’s answer was about localisation and product, not about winning on price. Credible, capital-hungry, and the main reason sales and marketing rose to 24% of revenue in FY2025 from 22%.

(3) Occasion expansion — real but small. Entertainment-driven travel +74% (Q1 2026), silver-generation Old Friends Club hotel bookings +100%, small-group tours +27% with 55% higher per-capita spend. These are genuinely incremental travel occasions rather than share shifts, which makes them high-quality growth. They are not yet large enough to move the group.

(4) AI as an efficiency lever — under-appreciated. With 16,718 customer-service employees, AI deflection of service contacts is a material margin opportunity that requires no revenue growth at all. Management has not quantified it.

Verdict: high-quality growth — organic, mix-improving, and increasingly international — decelerating from recovery rates toward a structural rate, and now interrupted by a self-inflicted compliance reset. The forward opportunity set is genuinely good and the inbound asset is differentiated. The material caveat is that the growth is being bought: sales and marketing grew 25% against revenue’s 17% in Q1 2026, and adjusted S&M rose 200bp as a share of revenue in FY2025. Growth this expensive is worth less than growth that isn’t.

6. Financial Quality

6.1 The earnings distortion — read this before any multiple

FY2025 reported net income was RMB 33.294 billion on revenue of RMB 62.409 billion — a 53.3% net margin, which no travel intermediary on earth earns. Operating income was RMB 15.8 billion. The gap is “other income” of RMB 21.3 billion, which exceeded operating income, against RMB 2.2 billion the prior year. Its composition, from the 20-F:

The RMB 15.4 billion is identified in the 20-F: in June 2025 Trip.com agreed to sell 34,372,221 MakeMyTrip Class B ordinary shares back to MakeMyTrip for ~US$3.0 billion for cancellation; the transaction closed in July 2025 and “we recorded a gain from investment of RMB15.2 billion.”

Here is the trap. Trip.com’s non-GAAP bridge removes share-based compensation and “fair value changes of equity securities investments and exchangeable senior notes.” A realised disposal gain is not a fair-value change, so it is not removed. The Q3 2025 reconciliation makes this unmistakable: against RMB 17.032 billion of other income, the non-GAAP bridge deducts only RMB 1.308 billion. Q3 2025 non-GAAP net income of RMB 19.156 billion — non-GAAP diluted EPS of RMB 27.56 — therefore contains the entire MakeMyTrip gain. The same contamination flows into FY2025 non-GAAP diluted EPS of RMB 45.59.

Management supplied the clean figure once, verbally, on the Q4 call: “Excluding gains from investments, net income attributable to Trip.com Group Limited was RMB 13.4 billion.”

Both the screen multiple and the company’s own non-GAAP multiple understate the true multiple by roughly 2.4×. Every quantitative screen in the market currently shows Trip.com at ~7× earnings. It is at ~16×.

This also invalidates one of our standard tools. The AZI own-history valuation index puts Trip.com’s P/E at the 0.9th percentile of its ten-year range — but that percentile is an artefact of inflated EPS and runs in the cheap direction. The P/S percentile of 7.7 and the P/B percentile of 23.2 are the reliable readings, and P/S is the cleanest: the stock is genuinely near its cheapest-ever on revenue.

6.2 Margins and the cost trajectory

Gross margin has been remarkably stable — 77.0% to 82.5% across an eleven-year window that includes a 49% revenue collapse. That stability is the clearest financial fingerprint of the asset-light model: cost of revenue is largely fulfilment and payment processing, and it scales.

Operating margin, company basis: 26.8% (2023) → 26.6% (2024) → 25.3% (2025). Q1 2026 was 24.3% against 25.8% in Q1 2025. Margins were already compressing before the compliance reset, and the reason is disclosed: adjusted sales and marketing rose to 24% of revenue in FY2025 from 22%, funding international expansion. In Q1 2026 every cost line outgrew revenue’s 17% except G&A: cost of revenue +23%, S&M +25%, product development +15%.

This matters for the bear case. It means the Q2 guidance is not the first sign of margin pressure — it is an acceleration of pressure that was already present for structural, self-chosen reasons.

6.3 Sizing the compliance hit

Q2 2025 revenue was RMB 14.843 billion. The +3–8% guide implies RMB 15.29–16.03 billion. Against Q1’s +17% trend, the shortfall is roughly 9–14 percentage points of growth ≈ RMB 1.3–2.1 billion per quarter ≈ RMB 5.3–8.3 billion annualised.

Management attributes it to two causes — macro (energy prices, airfares, geopolitical route disruption) and compliance. They did not quantify the split. ASSUMPTION: an even split, giving a compliance-related annualised revenue loss of RMB 2.7–4.2 billion. Because take-rate and value-added-service revenue carries near-100% incremental margin, that maps almost one-for-one to operating income: roughly 17–26% of FY2025 operating income of RMB 15.8 billion. We carry 15–30% as the working range.

The macro half deserves less weight than management gives it. Higher airfares reduce volumes but raise ticket values, and Trip.com’s air take is percentage-based — management themselves noted “part of the volume impact is offset by higher ticketing prices.” The compliance half has no such offset.

6.4 Cash flow — the genuine deterioration

Free cash flow fell 29% in a year revenue rose 17%. The proximate cause is working capital, specifically a RMB 5.68 billion increase in prepayments and other current assets, which the 20-F flags but does not explain. Prepayments to airlines and hotels for inventory access are the benign reading; a build of this size in a single year in a business with negative working capital warrants scrutiny, and the disclosure is thin. OPEN QUESTION.

The 0.43× cash conversion is arithmetically expected — net income contained RMB 19.4 billion of non-cash and investing gains that are correctly stripped out in the operating section. But even normalising for that, the working-capital drag is real and it is the reason the FCF yield is not as attractive as the earnings multiple suggests.

Capital intensity is genuinely trivial: capex of RMB 797 million on RMB 62.4 billion of revenue — 1.3%. Trip.com converts operating profit to cash without reinvestment, which is the strongest structural feature of the financial model.

6.5 Returns on capital — two answers, both true

ROIC.ai reports consolidated ROIC of 6.0% (2023) → 7.1% (2024) → 7.4% (2025). For a business with 80% gross margins and 25% operating margins, that looks like a contradiction. It is not.

The consolidated denominator contains RMB 62.2 billion of goodwill, RMB 13.0 billion of other intangibles (Qunar and Skyscanner, acquired a decade ago) and, at various points, RMB 32–38 billion of strategic equity stakes. Strip cash, held-to-maturity deposits and the investment portfolio, and net off non-debt current liabilities, and operating invested capital at 31 December 2025 is roughly RMB 71 billion — of which RMB 75.2 billion is goodwill and intangibles. In other words, invested capital in the operating business, excluding acquisition goodwill, is negative. The OTA runs on float: a cash conversion cycle of –446 days, RMB 18.9 billion of deferred revenue and RMB 20.0 billion of payables funding RMB 6.5 billion of property and equipment.

Both readings must be stated, and they answer different questions:

• Returns on tangible operating capital are effectively unbounded. That is the verdict on the business, and it is outstanding.
• Consolidated ROIC of ~7.4% is the verdict on the holding company — the price paid for Qunar and Skyscanner a decade ago, and the decision to hold a multi-billion-dollar marked-to-market equity portfolio inside an operating company. That is a capital-allocation verdict, and it is poor.

One caution: ROIC.ai reports FY2025 return on common equity of 54.6%, which is not reproducible — RMB 33.294 billion over average equity of ~RMB 156.75 billion is 21.2%, and even that is an artefact of the MakeMyTrip gain. The aggregator’s ROE is not usable for this issuer.

6.6 Balance sheet

At 31 March 2026:

A critical adjustment most screens miss: the balance-sheet “Investments” line of RMB 54.79 billion includes RMB 22.95 billion of held-to-maturity time deposits and financial products. The genuine strategic equity portfolio — Tongcheng Travel (0780.HK), residual MakeMyTrip, H World, BTG/Homeinns and others — is therefore RMB 31.84 billion (US$4.62 billion), not RMB 54.79 billion. It fell ~RMB 6.6 billion in Q1 2026 alone, a reminder that this is a marked-to-market equity book, not a cash-equivalent, and that its contribution to reported earnings is volatile in both directions.

Net cash of US$10.54 billion equals 38% of market capitalisation. Two qualifications keep that from being taken at face value. First, a meaningful portion is customer and supplier float: with a –446-day cash conversion cycle, roughly RMB 18 billion of the balance is working capital that must stay in the business. Second, most of it is onshore RMB, and repatriation to the Cayman holdco attracts a 5% withholding tax under the China–Hong Kong arrangement (PRC subsidiaries distributed RMB 14.7 billion to Ctrip.com (Hong Kong) in 2025, subject to that 5%). It is real cash, but it is not free cash.

Debt maturities are dateable and worth flagging. US$500 million of 1.50% exchangeable senior notes due 2027, and US$1.5 billion of 0.75% convertible senior notes due 2029 carrying an investor put on 15 June 2027 at par, with an effective conversion rate of 15.1193 ADS per US$1,000 — a conversion price of ~US$66.14. At US$43.64 the converts are ~34% out of the money, so the June 2027 put should be assumed exercised: ~US$2.0 billion of 2027 cash calls against US$10.5 billion of net cash. Comfortable, but it is a real constraint on how much of the US$5 billion repurchase authorisation can be deployed before mid-2027.

Verdict: economics that improve dramatically with scale, inside a holding company that dilutes them. The operating business is a negative-capital, 80%-gross-margin, 1.3%-capex float machine with genuinely outstanding tangible returns and a fortress balance sheet. The reported financials obscure all of it — a one-time disposal gain inflating earnings by 2.4×, a marked-to-market equity portfolio adding volatility, a decade-old goodwill block dragging consolidated ROIC to 7.4%, and a 29% free-cash-flow decline in a growth year that the disclosure does not adequately explain. Excellent business; opaque and deteriorating reported financials.

7. Capital Allocation

7.1 The record

In November 2023 the board adopted a “regular capital return policy” of discretionary annual buybacks and/or dividends. In February 2025 it declared the company’s first-ever dividend (US$0.30/ADS, ~US$200 million). Repurchases were 6.0 million ADS for US$300 million in 2024 and 10.0 million ADS for US$613 million in 2025. In August 2025 the board authorised a new repurchase programme of up to US$5.0 billion — roughly 18% of the current market capitalisation.

7.2 The Q1 2026 buyback — the strongest management signal in the file

Ordinary shares outstanding fell from 649,583,574 (31 Dec 2025) to 629,705,222 (31 Mar 2026) — 19.88 million shares, –3.1% in a single quarter. Over the same quarter, total Trip.com shareholders’ equity fell from RMB 170.818 billion to RMB 165.000 billion, a RMB 5.8 billion decline despite RMB 2.5 billion of net income.

Those two facts together imply roughly RMB 7–8 billion (~US$1.0–1.2 billion) of repurchases in Q1 2026 — nearly double the entire FY2025 programme, in one quarter, executed at roughly US$50–63 per ADS into the teeth of the post-SAMR decline. (ASSUMPTION: the quarterly repurchase amount is not separately disclosed in the Q1 release; this is inferred from the share count and the equity roll-forward.)

This deserves real weight. The best-informed buyer of the security, holding a US$5 billion authorisation and US$10.5 billion of net cash, doubled its purchase rate immediately after a regulatory shock. It is the mirror image of the “buyback that quietly stops” signal, and it is a revealed valuation opinion worth more than any multiple we can compute.

7.3 M&A and the investment portfolio

There has been no material acquisition since Skyscanner in December 2016 — a decade of organic discipline that should be credited, particularly against a Chinese platform peer group notorious for empire-building.

The offsetting story is the investment portfolio. The June 2025 MakeMyTrip sell-down — 34.4 million shares back to the issuer for US$3.0 billion, described in the 20-F as “part of our efforts to optimize our investment portfolio and enhance shareholder returns” — was genuinely good capital allocation: monetising an appreciated minority stake at scale into an issuer buyback, then recycling the proceeds into a US$5 billion authorisation for its own shares. That is textbook.

But the residual RMB 31.8 billion (US$4.6 billion) equity book remains, and it is 17% of market capitalisation held in marked-to-market stakes in Tongcheng (a direct competitor in which Executive Chairman James Liang also sits as a director), H World, BTG/Homeinns and others. It fell RMB 6.6 billion in Q1 2026. It is the primary reason consolidated ROIC screens at 7.4%. The strategic rationale — ecosystem alignment with hotel groups and a competitor-partner — is coherent, and the related-party disclosure is clean (Tongcheng paid Trip.com RMB 300 million and received RMB 61 million in FY2025). But shareholders own an operating company that also runs a US$4.6 billion equity fund, and they did not ask for the second one.

7.4 Incentives — the weakest link

At 31 March 2026, directors and officers as a group beneficially owned 63,766,293 ordinary shares (~9.7%). The composition matters far more than the total:

Alignment is overwhelmingly option-based, not ownership-based. The founder’s genuine equity stake is ~1.5% of shares outstanding against ~34 million options. Options are asymmetric: they pay for price appreciation and volatility and cost nothing on the downside. For a company whose central open question is whether management defends near-term monetisation or accepts a permanent take-rate reset in exchange for regulatory peace, an option-dominated incentive structure is not obviously aligned with a long-term owner. It is a legitimate concern, not a disqualifying one — Liang founded the company in 1999 and has run it through three cycles — but it should be named.

Share-based compensation was RMB 2.27 billion in FY2025 — 14.4% of company-basis operating income. Diluted shares (681.7 million, Q1 2026) exceed basic (649.0 million) by 32.7 million, a 5.0% overhang that includes the 2029 converts.

7.5 Insider transactions — read the Form 144s, not the Form 4s

A structural caveat that must be stated first: foreign private issuers are exempt from Section 16 reporting under Exchange Act Rule 3a12-3(b). Trip.com’s five-year EDGAR corpus contains exactly one Form 4. That is an artefact of FPI status, not evidence of a clean insider record, and reading it as the latter is a mistake.

The usable window is the Form 144 corpus (33 filings over five years). It contains a cluster that is difficult to ignore:

~US$106.8 million sold across three business days, at roughly US$71–74 per ADS — five days after the +14.9% Q2 2025 earnings pop (28 Aug 2025, $65.29 → $75.03), within ~6–10% of the eventual all-time high of $78.96, and roughly 65% above today’s price. The largest single line is 1,000,000 ADS of founder-family “Founder Shares” — not an option exercise, but a disposal of original equity. There is no offsetting record of discretionary open-market purchase by any insider anywhere in the five-year corpus.

Verdict: genuinely improved and, in the last two quarters, actively good — but with real structural blemishes. The positives are substantial and should not be understated: a decade without a value-destroying acquisition, an excellent US$3.0 billion monetisation of the MakeMyTrip stake, a first dividend, a US$5 billion authorisation, and an inferred ~US$1.0–1.2 billion of repurchases executed into the company’s own regulatory crisis — the single best piece of management-behaviour evidence available. Against that: a US$4.6 billion marked-to-market equity portfolio shareholders did not ask for and which caps consolidated ROIC at 7.4%; an incentive structure in which the founder holds ~1.5% of the equity and ~34 million options; a 5.0% dilution overhang; and insiders, including founder-family shares, selling ~US$107 million within 10% of the top four months before the SAMR disclosure. Capital allocation is now a modest positive for the thesis. Incentive alignment is not.

8. Changes and Headwinds — Last Two Years

The pattern is unambiguous. Everything that went right — the MakeMyTrip monetisation, the US$5 billion authorisation, the aggressive Q1 buyback, the 60–65% international growth — is a capital and franchise story. Everything that went wrong is a Chinese regulatory story, and it arrived in a tight, escalating sequence: complaints in November 2025, formal investigation in January 2026, the tool surrendered in March 2026, the rail summons in June 2026, and the guidance reset twelve days after the summons.

Two features of that sequence are analytically important. First, the guidance cut came twelve days after the rail summons, which strongly suggests the rail action — not the anti-monopoly probe alone — was the proximate trigger for quantifying the Q2 hit. Second, the company gave up the price-adjustment tool voluntarily and early (10 March), before any SAMR finding. That is the Meituan playbook: Meituan’s fine came in at ~3% of domestic revenue against Alibaba’s 4%, explicitly because of more cooperative rectification. Early surrender is expensive operationally but it is the correct strategy for minimising the penalty — and it is evidence that management expects an adverse finding.

Verdict: the last two years have strengthened the balance sheet and the capital-return posture while materially weakening the domestic earnings power and, more importantly, the predictability of it. The franchise is intact — 17% revenue growth in Q1 2026 with international compounding at 65% is not a broken business. What has been damaged is the monetisation model in China and the market’s ability to underwrite it. On net these changes weaken the thesis, and the weakening is permanent rather than cyclical, because a surrendered pricing tool and a rectified value-added-service model do not come back.

9. Risk Analysis

The risk matrix has an unusual shape and that shape is the investment case. The high-likelihood risks (#1, #2, #10) are largely known and partly priced; the highest-impact tail (#8, VIE) is genuinely smaller here than for the China-ADR complex the stock trades with; and the risk the market appears most focused on (#6, the fine) is the one we assess as low impact. The genuinely dangerous risk is #3 — that mid-single-digit growth is the new structural rate rather than a one-quarter rebasing — and it is unresolvable before the Q2 print.

10. Valuation Discussion

No price target and no recommendation appear in this section. This is an embedded-expectations analysis.

10.1 Rebuilding the capital structure by hand

Aggregator enterprise values for this issuer are unusable — market capitalisation is struck at fiscal-period end, and the “Investments” line conflates deposits with equity stakes. Rebuilt at the 24 July 2026 close:

Other reference points: P/B 1.14×, P/TBV 2.07×, FCF yield on market capitalisation 9.5% (3-year average) or 7.2% (FY2025 actual), net cash 38% of market capitalisation.

10.2 Own-history context

Today’s 7.2× is below every year-end and every intra-year low in the retrievable eleven-year series. The AZI own-history index corroborates on the clean metrics: composite 10.6th percentile, P/S 7.7th percentile, P/B 23.2nd percentile. (The 0.9th-percentile P/E is discarded for the earnings-distortion reason set out in the Financial Quality section.)

10.3 Is it a sector-wide de-rating? No.

Running the same own-history screen across the OTA complex at the same date (24 July 2026):

Every OTA peer sits mid-range in its own history; Trip.com sits at the 10th percentile. The de-rating is idiosyncratic. This is decisive for how the security should be understood: it is not a travel-cycle call, it is a China-regulatory call. Buying it is not a view that travel is cheap — it is a view that the Chinese regulatory outcome is less bad than 7.2× EBIT implies.

On honest earnings, Trip.com trades at ~15.9× against BKNG 23.4×, EXPE 22.9× and ABNB 34.9× — a 30–55% discount — while growing revenue faster than all three and holding net cash equal to 38% of its market capitalisation. (P/S is not the right cross-sectional metric here: TCOM’s 3.67× against Expedia’s 2.20× reflects TCOM’s ~25% operating margin against Expedia’s ~14%.)

10.4 Embedded expectations — what the price requires

At an enterprise value of US$16.94 billion against trailing operating income of US$2.345 billion, the market pays 7.22× EBIT.

Applying 12× EV/EBIT — the low end of Trip.com’s own eleven-year range, not a peer multiple — to today’s enterprise value implies normalised EBIT of US$1.41 billion: a 40% permanent decline from the trailing level. On core enterprise value (stakes at book), 12× implies US$1.03 billion — a 56% permanent decline.

Set that against the sizing of the actual damage above: a permanent compliance cost of 15–30% of operating income. The market is discounting between 1.3× and 3.7× the damage that can be constructed from the disclosure — and it is doing so with no credit for a 200 million-inbound-traveller opportunity growing at 90%, an international platform compounding at 60–65%, or a US$5 billion repurchase authorisation being drawn at ~US$1 billion a quarter.

The counter-argument, which must be stated fairly: the market is not pricing the known damage, it is pricing the variance. SAMR has not ruled. No reserve exists. The company cannot estimate a loss. Management has given exactly one quarter of post-compliance guidance and called the deceleration “ordinary normalization.” In that information state a wide discount is rational, and the multiple should be expected to stay compressed until the Q2 and Q3 prints resolve the level-versus-rate question.

10.5 Scenario analysis

All scenarios use normalised EBIT and multiples anchored to Trip.com’s own 10–25× eleven-year range, not to peers. Assumption-heavy by construction.

The distribution is positively skewed from US$43.64 — roughly –15% to the bear case against +21% to the base and +70% to the bull. That skew, not the headline multiple, is the honest quantitative case.

What the market is underwriting correctly: that a permanent monetisation reset is underway; that management cannot yet quantify it; that a Chinese platform’s returns are subject to policy in a way a Western platform’s are not; and that the strategic-stake portfolio is volatile and should not be capitalised at par.

What it may be underwriting incorrectly: conflating the fine (~2% of market cap, affordable) with the remedy; extrapolating one guided quarter into a permanent growth rate; capitalising a business at 7× EBIT while its fastest-growing segment compounds at 60–65%; ignoring that a 5.7% blended take rate leaves far less to confiscate than a 14.5% one; and pricing the security as a China-internet ADR rather than on travel fundamentals.

11. Variant Perception

11.1 The consensus

Trip.com is a Chinese platform monopolist that got caught. Its regulator will extract a fine and force it to give up the pricing practices that made it profitable; growth has collapsed from 17% to 3–8% and will not recover; the earnings multiple is optically low but the earnings are about to fall; and no Western investor can underwrite a Chinese regulatory outcome. Sell-side sentiment reflects this — Seeking Alpha downgraded to Hold on 29 June specifically citing “Q2 revenue guidance miss and ongoing antimonopoly investigation,” warning of “substantial fines and margin compression.”

11.2 The strongest bull case

Four legs, in descending order of strength.

(1) The multiple everyone is quoting is wrong in the cheap direction, and the multiple that is actually right is still cheap. The screen says 7×; the truth is ~15.9× ex-investment-gains, or 13.6× adding back SBC. That correction should reduce enthusiasm — and yet 15.9× on a business growing revenue 17%, with 38% of its market cap in net cash and a 9.5% three-year-average free-cash-flow yield, against Booking at 23.4× and Airbnb at 34.9×, is still a 30–55% discount. Correcting the error strengthens the argument’s credibility without destroying the conclusion.

(2) The take-rate asymmetry. Trip.com earns 5.7% of gross bookings. Booking earns 14.5%, Airbnb 13.4%, Expedia 12.3%. The company under investigation for monopolistic abuse is the least extractive major OTA in the world by more than 2×. SAMR’s case is about conduct, not price level, so this is no legal defence — but it is a hard economic constraint on how much a remedy can cost. You cannot confiscate a rent that was never collected. Consensus is implicitly applying an Alibaba-shaped remedy (where “choose one of two” exclusivity was worth an enormous, permanent margin) to a platform with less than half the toll and no equivalent exclusivity mechanism.

(3) The fine and the remedy are being conflated. On the Alibaba (4% of China revenue) and Meituan (~3%) precedents, a fine is RMB 1.1–1.5 billion, ~US$160–215 million, ~2% of market capitalisation — about six weeks of gross cash. Alibaba’s RMB 18.2 billion fine was a one-day event; what permanently impaired Alibaba was losing merchant exclusivity. Here the analogous permanent loss — the AI price-adjustment tool — has already been surrendered, voluntarily, on 10 March 2026, and its cost is already inside the Q2 guide. The largest identifiable permanent damage is therefore behind the disclosure, not ahead of it.

(4) The market is not pricing this as a travel company at all. The factor evidence is unusually clean: the loading on the “Travel Leisure Giants” basket is +0.010 — indistinguishable from zero — while Country: China is +0.935 and Industry: Social Media is +1.089. The factor-similar peer list is Xiaomi, Meituan, Baidu, 360 Finance, ZTO, Vipshop, Yum China, JD, PDD — with no Western travel name anywhere in it. The multiple is being set by the China-ADR risk premium, not by travel fundamentals, which is precisely why the own-history percentile is 10.6 while every OTA peer sits at 35–64. If the regulatory question resolves, the security has to be re-underwritten as a travel asset, and that is a different multiple.

11.3 The strongest bear case

Also four legs, and they are serious.

(1) It is a rate reset, not a level reset. Management’s framing — “ordinary normalization,” “constructive for the longer term” — is what a company says when growth is structurally lower. If the compliance rectification permanently removes the value-added-service layer across rail, hotel ranking and promotional mechanics, then mid-single-digit domestic growth is the new normal and 7.2× EBIT on a shrinking domestic business is not cheap. There is exactly one quarter of guidance and no post-compliance actual result.

(2) A 5.7% take rate cuts both ways. The bull frames it as “little left to confiscate.” The bear frames it as “no cushion.” A platform earning 14.5% can surrender 3 points and still earn 11.5%. A platform earning 5.7% that surrenders 1.5 points loses 26% of its revenue. Trip.com’s thin toll makes it more fragile to a monetisation reset in percentage terms, not less. Both readings are logically available and the truth depends entirely on the remedy’s absolute size — which is unknowable today.

(3) The financial quality is deteriorating beneath the headline. Free cash flow fell 29% in a 17% growth year on an unexplained RMB 5.68 billion prepayments build. Operating margin was already compressing (26.8% → 26.6% → 25.3% → 24.3% in Q1 2026) before compliance, because international expansion is being bought with sales and marketing rising 200bp to 24% of revenue. Consolidated ROIC is 7.4%. SBC is 14.4% of operating income. None of that is a crisis; all of it argues against paying up for a re-rating.

(4) The insider record and the incentive structure. Insiders including founder-family shares sold ~US$106.8 million within three days, five days after the earnings pop, within 10% of the all-time high, four months before the SAMR disclosure and roughly 65% above today’s price. The founder’s genuine equity stake is ~1.5% against ~34 million options. The people with the most information monetised at the top, and the people setting strategy are paid in instruments that reward volatility.

11.4 The 3–5 assumptions that actually decide it

11.5 Where the tape says consensus may be offsides

The positioning read is genuinely informative here rather than decorative. Trip.com is a quantitatively confirmed falling knife: 3-month –54%, 6-month –52%, 1-year –32% (all annualised), 1-year Sharpe –0.90, 1-year maximum drawdown –49.5%, trading below all three major moving averages, with a ten-year annualised return of +0.17% and a ten-year Sharpe of –0.04. Nothing about the trend is supportive, and anyone framing this as a contrarian entry must own that the ten-year holder of this security has earned nothing.

But the composition of the risk is where consensus looks vulnerable. Idiosyncratic volatility is 34.6% annualised and the factor model’s R² is only 0.328 — meaning roughly two-thirds of this security’s variance is company-specific, not China-factor. The market is applying a China-ADR discount rate to a security whose returns are mostly not explained by China-ADR factors. That is the mispricing, if there is one: the discount is being set by the correlation and the outcome will be set by the idiosyncrasy.

Verdict: a genuine variant perception exists, and it is narrower and more specific than “China is cheap.” It is this — the market has priced a 40% permanent destruction of operating income into a platform whose entire blended toll is 5.7%, whose most damaging remedy has already been voluntarily surrendered, whose fine exposure is ~2% of market capitalisation, and whose fastest-growing segment is compounding at 60–90% with explicit state support. The offsetting reality is that the decisive question — level reset or rate reset — is genuinely unanswerable before the Q2 print, and the price action gives no reason to front-run it.

12. Fact vs. Interpretation

13. Open Questions

1. What is the rail-specific share of the RMB 8.2 billion of non-air transportation revenue? The company has never disaggregated rail from bus, ferry, ground transport and international ground. Without it, the rail-crackdown exposure can only be bounded, not sized.
2. What is the split of the Q2 deceleration between macro and compliance? Management named both causes and quantified neither. The entire compliance-cost estimate in this article rests on an assumed even split.
3. What drove the RMB 5.68 billion increase in prepayments and other current assets in FY2025? The 20-F identifies it as the cause of a 29% FCF decline and does not explain it. Supplier prepayments for inventory access is the benign reading; it is not the only one.
4. What was the actual Q1 2026 repurchase amount? Inferred at ~US$1.0–1.2 billion from the share count and equity roll-forward; not disclosed. The inference underpins the strongest management-behaviour argument in this memo.
5. What is the international segment’s standalone profitability? Management says the margin profile has “significantly improved” and is “structurally stronger” but has never disclosed an international segment margin. With international at ~40% of revenue and growing at 60–65%, this is the single largest disclosure gap.
6. Will SAMR’s remedy extend beyond the surrendered pricing tool? Commission caps, mandated ranking neutrality, or forced supplier-pricing autonomy would each be materially worse than the fine.
7. Is there any timetable for the SAMR decision? Alibaba’s ran ~4 months, Meituan’s ~5 months. Trip.com’s is now past 6 months with no disclosed milestone.
8. What is the accounting policy and independent-valuation basis for the RMB 31.8 billion strategic portfolio? It moved RMB 6.6 billion in one quarter and flows through earnings.
9. Why does Trip.com consolidate a Qunar entity that is 57% owned by outside investment vehicles, and what are the economics of that minority to Trip.com shareholders?
10. Government grants were RMB 1.3 billion in FY2025 (8% of operating income), up from RMB 787 million. What are they, and are they durable?

14. What Must Be True

Bull case — what must be true

1. The Q2 2026 deceleration is a level rebasing, not a rate change. Growth resumes at ≥10% once the compliance adjustments annualise.
2. SAMR’s remedy is limited to conduct already surrendered, plus a fine at or below the Meituan precedent (~3% of domestic revenue, ~RMB 1.1bn).
3. The international platform keeps compounding at ≥40–50% and inflects to segment profitability, carrying group growth while China rebases.
4. The FY2025 working-capital drag reverses, restoring free cash flow toward RMB 19–21 billion.
5. Management continues repurchasing at scale, retiring >10% of shares against the US$5 billion authorisation before the June 2027 convert put.

Falsification test for the bull case: Q3 2026 revenue guidance issued in the low single digits, or full-year 2026 revenue growth guided below 8%. That would establish the reset as a permanent rate change rather than a one-quarter rebasing, at which point 7.2× EV/EBIT is not a discount — it is the correct multiple for a structurally decelerating, regulator-constrained domestic franchise. A second, independent falsifier: international gross-bookings growth printing below 40%, which would remove the offset that makes the group arithmetic work.

Bear case — what must be true

1. The compliance rectification permanently removes a large share of Chinese monetisation — the pricing tool, rail value-added services, and promotional/ranking mechanics — costing ≥30% of operating income.
2. SAMR imposes structural remedies beyond conduct, such as commission caps or mandated supplier pricing autonomy.
3. Super-app competitors take domestic share while Trip.com is constrained, so the rebased level erodes further rather than stabilising.
4. The margin compression already underway continues, as international expansion requires permanently higher sales and marketing without a profitability inflection.
5. The strategic portfolio and working-capital drag persist, keeping consolidated returns near 7% and free cash flow below RMB 15 billion.

Falsification test for the bear case: SAMR concludes with a fine at or below ~RMB 1.5 billion and no structural remedy beyond the already-surrendered pricing tool, accompanied by two consecutive quarters of revenue growth above 10%. That combination would establish that the permanent damage was smaller than the price implies and already disclosed, and that the franchise absorbed the reset — leaving a business compounding at low-double-digits with a 5.7% toll, 38% of its market cap in cash, and a US$5 billion buyback running. A second, independent falsifier: disclosure of international segment operating profitability at scale, which would prove the 40%-of-revenue growth engine is accretive rather than a margin drain.

15. Source Appendix

See the separate Source Appendix (Appendix B) for the full source list with URLs and access dates.

This article contains no investment recommendation and no price target outside the clearly-labelled Claude's Take block, which is the author’s own subjective view. Nothing herein should be construed as investment advice. The author may or may not hold a position in the securities discussed. Readers should conduct their own research and consult a qualified adviser before making any investment decision.

APPENDIX A — Standard Diligence Questionnaire

Report date: 25 July 2026 · Reference price: US$43.64 · RMB 6.8905 = US$1.00

A structured diligence checklist applied to the company. Fact / Interpretation / Assumption labels are applied where the distinction is material.

General

What thoughtful questions have other investors asked about this company?

The sell-side Q&A across the last two calls converges on five questions, and their quality is uneven.

The best question asked was Joyce Ju’s (Bank of America) on the Q1 2026 call: given that value-added services are closely linked to the monetisation of rail ticketing, how should the financial impact be understood? This is the only question on any call that correctly identified the mechanism by which regulation converts into lost revenue — that Chinese rail tickets carry no commission and all the economics sit in the value-added layer regulators just attacked. The CFO’s answer was evasive in a specific and informative way: she said rail’s “direct contribution to overall revenue and earnings has meaningfully declined over the years” without ever giving a number. FACT: non-air transportation is RMB 8.2 billion, or 13.1% of group revenue (RMB 22.5bn transportation less RMB 14.3bn air). That is not small, and nobody made her say it.

Wei Fang (Mizuho) asked for an update on the regulatory review and its impact — answered with process language (“fully cooperating,” “premature to speculate”) and the admission that adjustments “may introduce some near-term impacts.”

Alex Yao (JPMorgan) and Simon Cheung (Goldman Sachs) both asked about AI agent disintermediation. James Liang’s answer is the most structurally coherent offered by any OTA management: the OTA rests on inspiration / transaction / service; AI agents excel at inspiration, which is the least defensible layer, and their rise therefore reinforces the importance of the transaction and service layers. Trip.com is simultaneously building vertical travel models and opening agent-to-agent transaction rails.

Parash Jain (HSBC) asked the sharpest competitive question — Agoda in APAC — and received a localisation answer with no share or margin data.

The question nobody asked, on either call: what is the international segment’s standalone operating margin? International is ~40% of revenue growing at 60–65%, management repeatedly asserts its profitability is “structurally stronger,” and it has never been disclosed. That is the largest disclosure gap in the file.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low?

Neither, and the question is being obscured by an accounting artefact. FACT: FY2025 reported net income of RMB 33.3 billion is a record — but RMB 21.3 billion of it was “other income,” including a RMB 15.4 billion gain on the MakeMyTrip disposal. Management’s own ex-investment-gains figure is RMB 13.4 billion. On that basis earnings are neither at a high nor a low; they are at a structurally rebased level that is about to fall, because Q2 2026 revenue is guided to +3–8% from +17% with explicit margin compression.

INTERPRETATION: the cleanest way to read the cycle is operating margin: 26.8% (2023) → 26.6% (2024) → 25.3% (2025) → 24.3% (Q1 2026). Earnings quality has been declining for three years, driven first by deliberate international investment and now by compliance.

Driven by the external environment or internal actions?

Unusually, mostly internal — but not voluntarily. The Q2 deceleration is attributed to two causes: external macro (energy prices, airfares, geopolitical route disruption) and internal “operational adjustments… to align with evolving industry standards and compliance frameworks.” The second is a self-executed response to regulatory pressure, which makes it internal in mechanism and external in cause. The macro half deserves less weight: higher airfares cut volumes but raise ticket values, and Trip.com’s take is percentage-based — management conceded “part of the volume impact is offset by higher ticketing prices.”

How stable are revenues?

Highly stable in mix, violently unstable in level under exogenous shock. FACT: revenue went RMB 35.7bn (2019) → 18.3bn (2020) → 20.0bn (2021) → 20.0bn (2022) → 44.5bn (2023) → 53.3bn (2024) → 62.4bn (2025). A 49% collapse and a 122% recovery in five years. But gross margin never left the 77.0–82.5% band across the entire cycle, which is the strongest evidence that the cost structure is genuinely variable and the franchise does not break under stress.

Outlook for products/services?

Divergent by segment, and the divergence is the story:

• Accommodation (41.8% of revenue, 9.3% take): +21% in FY2025, +17% in Q1 2026. The profit engine. Take rate under regulatory pressure.
• Air (22.9% of revenue, 2.6% take): +11–12%. A volume and customer-acquisition business, not a profit pool.
• Non-air transportation (13.1% of revenue): directly targeted by the June 2026 tri-agency rail action. The end-state visible in the official 12306 channel is zero fees.
• International (~40% of revenue): +60% FY2025, +65% Q1 2026. The growth engine and the only segment with genuine multi-year runway.

How big will this market be — growing, shrinking, domestic or international?

Growing, and the mix is shifting international. FACT: international rose from ~35% to ~40% of revenue in one year. FACT: inbound bookings +90% year-over-year in Q1 2026, ~20 million inbound travellers served in 2025, management targeting 200 million over five years. The framing management uses is verifiable: inbound tourism is ~0.5% of China’s GDP against >10% for Thailand and 5–6% for France, Italy and Spain. INTERPRETATION: a 5–10× gap is real, the policy support is concrete (80+ visa-free countries, 10-day visa-free transit, foreign-card acceptance on Alipay/WeChat Pay), and Trip.com is uniquely positioned because it owns both Chinese supply relationships and international demand channels (Trip.com, Skyscanner) — a combination neither Booking nor Meituan has.

Business Quality & Competitive Moat

Is the industry getting more or less competitive?

More — and from an unexpected direction. The threat is not new OTA entrants; there has been no scaled new entrant in a decade because the entry economics are unattractive (you must replicate supply integration and 24/7 multilingual fulfilment to earn a 5.7% toll). The threat is adjacent incumbents redeploying existing demand assets: Meituan from local services, Douyin from short-video commerce, Fliggy from e-commerce. Each already owns the customer relationship Trip.com must acquire.

INTERPRETATION: the more consequential change is that the regulator is now compressing returns directly. Marathon’s capital cycle predicts high returns attract capital and mean-revert. Here returns are being reduced by decree, which will not mean-revert on its own.

How profitable is the business (ROIC, ROE)?

This requires two answers, and giving only one is misleading.

FACT: consolidated ROIC is 7.4% (2025), 7.1% (2024), 6.0% (2023) — poor, below any plausible cost of capital.

INTERPRETATION: that denominator contains RMB 62.2bn of goodwill and RMB 13.0bn of intangibles from Qunar (2015) and Skyscanner (2016), plus a RMB 31.8bn strategic equity portfolio. Strip cash, held-to-maturity deposits and stakes, and net off non-debt current liabilities, and operating invested capital is roughly RMB 71 billion — of which RMB 75.2 billion is goodwill and intangibles. Invested capital in the operating business, excluding acquisition goodwill, is negative. The OTA runs on float: a –446-day cash conversion cycle, RMB 18.9bn of deferred revenue and RMB 20.0bn of payables funding RMB 6.5bn of property and equipment, with capex at 1.3% of revenue.

So: returns on tangible operating capital are effectively unbounded; consolidated ROIC of 7.4% is a verdict on a decade-old acquisition and an equity portfolio, not on the business. Both are true and they answer different questions.

Caution: ROIC.ai reports FY2025 ROE of 54.6%, which is not reproducible (RMB 33.294bn / ~RMB 156.75bn average equity = 21.2%) and is in any case an artefact of the MakeMyTrip gain. Not usable.

How profitable is the industry — how many competitors, what barriers to entry?

The Chinese online travel industry supports one dominant player (Trip.com, ~48% of online hotel bookings), one scaled challenger with a superior demand asset (Meituan), one fast-growing content-led entrant (Douyin), one e-commerce affiliate (Fliggy) and one listed partner-competitor Trip.com part-owns (Tongcheng). Barriers to entry are real but low-yielding: supply integration, fulfilment infrastructure and 24/7 multilingual service are expensive and slow to build, but the reward is a 5.7% toll.

INTERPRETATION — the Greenwald test. Trip.com’s ~48% online hotel share has been broadly stable for several years despite sustained super-app attack, which passes the market-share-stability test on the domestic hotel franchise. But a firm with an unconstrained barrier converts it into price, and Trip.com earns less than 40% of Booking’s toll. The resolution: the barrier is real, but industry structure caps what it can extract — administered transport pricing, fragmented low-ADR hotel supply, and super-app cross-subsidy. The moat is genuine and capped.

Can the business be easily understood?

Yes. It is a toll booth: aggregate fragmented supply, aggregate fragmented demand, take a percentage. The complications are jurisdictional (VIE structure, RMB reporting, FPI disclosure) and accounting (a disposal gain inside non-GAAP earnings, held-to-maturity deposits inside the “Investments” line), not economic.

Can it be undermined by foreign low-cost labour?

No — inverted, in fact. Trip.com’s 16,718 customer-service employees delivering 24/7 support in 30+ languages are a Chinese low-cost-labour advantage that Western OTAs cannot match at comparable cost. It is a genuine structural asset, particularly for the inbound strategy.

Do brands matter?

Moderately, and less than in the West. Trip.com runs four brands deliberately segmented — Ctrip (Chinese full-service), Qunar (Chinese value), Trip.com (international), Skyscanner (metasearch). INTERPRETATION: Chinese travel brand loyalty is weak; consumers multi-home and price-compare aggressively. Airbnb’s “brand as distribution channel” advantage (which lets it avoid the Google tax) has no Chinese analogue. Brand matters most in the international business, where Trip.com is building recognition from a low base against Agoda and Booking.com — and that is precisely where the sales-and-marketing spend is going (24% of revenue in FY2025, up from 22%).

What is the nature of competition?

Service, supply breadth and fulfilment reliability — per management, and it is at least partly true. CEO Jane Sun’s competitive answer on the Q4 call cited 24/7 call centres, comprehensive product, global coverage, and evacuating customers from conflict zones within two minutes. INTERPRETATION: these are real differentiators but they are operational excellence, not structural captivity — a competitor can hire call-centre staff. Note also that management asserted the domestic market “remains structurally rational” with competition centred on “quality of service, supply coverage, and user experience rather than aggressive price wars.” That is an assertion, not evidence, and firms in genuinely rational industries rarely need to say so.

Customers’ switching costs?

Near zero, and this is the moat’s weakest point. No contractual lock-in, no meaningful loyalty economics, apps are free, and Chinese travellers routinely price-compare across five platforms. What retains users is habit, breadth and service quality — genuine but not structural. This is why the AI price-adjustment tool mattered so much: guaranteeing the lowest price was the substitute for switching costs. It was surrendered on 10 March 2026.

Financial Condition & Balance Sheet

Assets not fully recognised on the balance sheet?

Three, all material:

1. The float. A –446-day cash conversion cycle means customers and suppliers fund the business. RMB 18.9bn of deferred revenue and RMB 20.0bn of payables are recorded as liabilities but function as permanent, interest-free capital. This is the single most valuable asset the company owns and it appears only as a liability.
2. The supply graph and integrations. Decades of direct technical connections to Chinese hotels, airlines and the rail system, plus ~450,000 hotels and ~180,000 attraction products globally. Fully expensed.
3. The strategic portfolio at book. RMB 31.8 billion of stakes in Tongcheng, MakeMyTrip (residual), H World and BTG/Homeinns are marked to market, so this is arguably fairly stated — but the MakeMyTrip disposal realised US$3.0 billion at a RMB 15.2 billion gain, which suggests carrying values had lagged.

Off-balance-sheet liabilities?

The SAMR contingency, and it is unquantified by design. The 20-F states the company “is unable to predict the timing, outcome or consequences of the investigation, or estimate the possible loss” and that findings “could result in a significant fine, other financial penalties and/or changes to the Company’s business practices and may have a material adverse effect.” No reserve has been recorded. On the Alibaba (4% of China revenue) and Meituan (~3%) precedents the fine is roughly RMB 1.1–1.5bn (~US$160–215m) — about 2% of market capitalisation. The unbooked liability that matters is not the fine; it is the permanent remedy.

Otherwise: modest operating leases (right-of-use asset RMB 835m), and US securities class actions (class period 30 Apr 2024 – 13 Jan 2026) that are immaterial against a US$10.5bn net cash position.

How conservative is the accounting?

Mixed, with one genuine problem.

Conservative: revenue recognised at or near travel rather than booking; costs largely expensed as incurred; the VIE disclosure is unusually clear (VIEs are only 5% of total assets).

Not conservative — the problem: management’s non-GAAP measure excludes fair-value changes on equity securities but not realised disposal gains. The Q3 2025 reconciliation deducts only RMB 1.308bn against RMB 17.032bn of other income, so non-GAAP EPS of RMB 27.56 contains the entire RMB 15.2bn MakeMyTrip gain, and FY2025 non-GAAP EPS of RMB 45.59 is contaminated the same way. This is not a rules violation — a realised disposal genuinely is not a fair-value change — but the effect is that the company’s own preferred metric overstates ongoing earnings by roughly 2.4×. The clean figure (RMB 13.4 billion) was given once, verbally, on a call, and appears in no press release, no reconciliation table and no filing headline.

Also worth watching: the RMB 5.68 billion FY2025 increase in prepayments and other current assets that drove a 29% free-cash-flow decline, flagged in the 20-F and not explained. And government grants of RMB 1.3 billion (8% of operating income), up from RMB 787 million, sitting in other income.

How CapEx-hungry is the business?

Barely at all — this is a defining strength. FY2025 capex was RMB 797 million on RMB 62.4 billion of revenue = 1.3%. Comparable to Booking (~1.1%) and better than Expedia (~3%). Operating profit converts to cash without reinvestment; growth is funded by sales and marketing (an expense) rather than capital.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy?

FACT: free cash flow RMB 21.4bn (2023) → 19.0bn (2024) → 13.6bn (2025) — a 29% decline in a year revenue rose 17%, caused by the prepayments build. Three-year average ≈ RMB 18.0bn (US$2.6bn), a 9.5% yield on market capitalisation and 15.4% on enterprise value.

Uses, and the trajectory is genuinely improving:

The philosophy is a November 2023 “regular capital return policy” of discretionary annual buybacks and/or dividends. FACT: August 2025 brought a US$5.0 billion repurchase authorisation (~18% of market capitalisation).

Significant acquisitions recently?

None since Skyscanner in December 2016 — a decade of organic discipline that deserves credit against a Chinese platform peer group known for empire-building. The notable transaction ran the other way: the June 2025 sale of 34,372,221 MakeMyTrip shares back to the issuer for ~US$3.0 billion, described in the 20-F as “part of our efforts to optimize our investment portfolio and enhance shareholder returns,” booking a RMB 15.2 billion gain. INTERPRETATION: monetising an appreciated minority stake at scale and recycling into a US$5 billion authorisation for one’s own shares is textbook good capital allocation.

Buying back shares?

Yes — and recently, aggressively. FY2024: 6.0m ADS / US$300m. FY2025: 10.0m ADS / US$613m. And then: FACT: shares outstanding fell from 649,583,574 (31 Dec 2025) to 629,705,222 (31 Mar 2026) — 19.88 million shares, –3.1% in one quarter — while total shareholders’ equity fell RMB 5.8 billion despite RMB 2.5 billion of net income.

ASSUMPTION: those two facts imply roughly RMB 7–8 billion (~US$1.0–1.2 billion) of Q1 2026 repurchases at ~US$50–63/ADS — nearly double the entire FY2025 programme, in a single quarter, executed into the company’s own regulatory crisis. The amount is not separately disclosed. INTERPRETATION: the best-informed buyer doubling its purchase rate immediately after a regulatory shock is a revealed valuation opinion worth more than any multiple.

One constraint: the US$1.5 billion 2029 convertibles carry an investor put on 15 June 2027 and are ~34% out of the money (strike ~US$66.14), so ~US$2.0 billion of 2027 cash calls (with the US$500m 2027 exchangeables) competes with the authorisation.

Issuing large amounts of new shares to insiders?

Yes, and this is the weakest part of the file. FACT: SBC was RMB 2.27 billion in FY2025 = 14.4% of company-basis operating income. Diluted shares (681.7m) exceed basic (649.0m) by 32.7 million — a 5.0% overhang. New option grants have run 11.4–12.3 million per year with four-year vesting.

Compensation policy of directors/management?

FACT: at 31 March 2026, directors and officers as a group held 63,766,293 shares (~9.7%) — but the composition is the finding:

INTERPRETATION: alignment is overwhelmingly option-based rather than ownership-based. The founder’s genuine equity stake is ~1.5% against ~34 million options. Options pay for appreciation and volatility and cost nothing on the downside. For a company whose central question is whether management defends near-term monetisation or accepts a permanent reset for regulatory peace, that is not obviously aligned with a long-term owner. It is a legitimate concern, not disqualifying — Liang founded the company in 1999 and has run it through three cycles.

Motivations of management?

Mixed evidence, honestly presented.

Supportive: a decade without a value-destroying acquisition; an excellent US$3.0bn portfolio monetisation; a first dividend; a US$5bn authorisation; ~US$1.0–1.2bn of buybacks into their own crisis; and voluntarily surrendering the AI pricing tool on 10 March 2026 before any SAMR finding — expensive operationally, but the correct strategy for minimising the penalty (Meituan’s ~3% fine versus Alibaba’s 4% was explicitly attributed to more cooperative rectification).

Concerning: FACT — insider sales. Note first that foreign private issuers are exempt from Section 16 (Exchange Act Rule 3a12-3(b)), so the single Form 4 in five years is a structural artefact, not a clean record. The Form 144 corpus is the usable window, and it shows:

~US$106.8 million across three business days at ~US$71–74 — five days after the +14.9% Q2-2025 earnings pop, within ~6–10% of the all-time high, four months before the SAMR disclosure, and roughly 65% above today’s price. The largest line is founder-family original equity, not an option exercise. No discretionary open-market purchase appears anywhere in the five-year corpus.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer?

An ADR — and the structural details matter. Trip.com Group Limited is a Cayman Islands exempted company; the NASDAQ-listed security is an American Depositary Share where 1 ADS = 1 ordinary share. It is a foreign private issuer: it reports on Form 20-F annually and Form 6-K episodically, not 10-K/10-Q; there are no quarterly SEC-mandated financial statements; and insiders are exempt from Section 16 reporting. It maintains a dual primary listing on the Hong Kong Stock Exchange (9961), which materially reduces forced-delisting risk relative to single-listed China ADRs. Financials are reported in RMB. Depositary fees apply to ADS holders. Not an MLP; no K-1.

VIE exposure — the key structural datum: the consolidated VIEs (Ctrip Commerce, Shanghai Huacheng, Qunar Beijing) constituted only 5% of total assets at 31 December 2025, far below the proportion at Alibaba, Baidu or JD. ADS holders own a Cayman holding company, not equity in the VIEs — the standard risk — but the great majority of the economics sit in directly-owned PRC subsidiaries.

Dividend policy?

FACT: no dividend was ever paid until 2025. In February 2025 the board declared the first: US$0.30 per ADS, ~US$200 million (RMB 1.423 billion paid). It is discretionary under the November 2023 capital-return policy, with the board explicitly reserving discretion over form, timing and amount. Payout ratio ~4% of reported net income. INTERPRETATION: the dividend is a signalling instrument, not an income proposition; buybacks are the primary return channel and are running roughly 3–5× the dividend.

Repatriation note: PRC subsidiaries distributed RMB 14.7 billion to the Hong Kong holding company in 2025, subject to 5% withholding under the China–Hong Kong arrangement. No subsidiary paid dividends directly to the Cayman parent in 2023, 2024 or 2025.

How profitable is the business?

FY2025: gross margin 80.6%, operating margin 25.3% (company basis), adjusted EBITDA margin ~30.3%. Reported net margin of 53.3% is meaningless — see below. Honest net margin, ex-investment gains: RMB 13.4bn / RMB 62.4bn = 21.5%.

Is net income diverging from cash from operations?

Yes, dramatically — and this is the most important quality-of-earnings question on the list.

FACT: FY2025 net income RMB 33.4 billion against cash from operations of RMB 14.4 billion — a ratio of 0.43×, versus 1.15× in 2024 and 2.22× in 2023.

INTERPRETATION: the divergence has two distinct causes and only one is benign.

1. Benign: RMB 19.4 billion of the net income was non-cash or investing gains (the RMB 15.4bn disposal gain plus RMB 4.0bn of fair-value marks), correctly stripped out in the operating section. This portion of the divergence is exactly what should happen.
2. Not benign: even after that adjustment, working capital consumed RMB 5.73 billion, of which RMB 5.68 billion was an increase in prepayments and other current assets. The 20-F identifies this as the driver of the FCF decline and does not explain it. In a business that historically runs on negative working capital, a build of this size in a single year warrants scrutiny.

The resulting valuation gap is material: the reported P/E is 6.4×, management’s non-GAAP P/E is 6.7×, and the honest ex-investment-gains P/E is ~15.9× (13.6× adding back SBC). Every screen in the market is currently showing this company at roughly 40% of its true multiple.

Risks & Downside

What factors would cause the stock to decline?

In descending order of expected impact:

1. Q3 2026 guidance also in the low single digits — proving the reset is a rate change, not a level rebasing. This is the single most dangerous outcome and is unresolvable before late August 2026.
2. A SAMR remedy extending beyond conduct — commission caps, mandated ranking neutrality, or forced supplier-pricing rules. Each would be materially worse than any fine.
3. International gross-bookings growth falling below 40%, removing the offset that makes the group arithmetic work while China rebases.
4. Continued FCF deterioration if the prepayments build proves structural.
5. Domestic share loss to Meituan/Douyin while Trip.com is constrained by rectification.
6. Further mark-downs on the RMB 31.8 billion strategic portfolio (it fell RMB 6.6 billion in Q1 2026 alone).
7. A broad China-ADR de-rating — to which this security is highly exposed given a Country: China factor loading of +0.935.

Risk of a catastrophic loss?

Low, and materially lower than the China-ADR complex it trades with. The reasons are specific:

• US$10.54 billion of net cash — 38% of market capitalisation. Even the punitive end of the statutory 1–10% AML fine band is affordable several times over.
• Debt is modest and well-termed: RMB 31.4 billion total, against RMB 104.0 billion of liquid assets. The 2027 obligations (~US$2.0 billion including the convert put) are covered five times over.
• VIEs are only 5% of total assets, and the HKEX dual primary listing provides a genuine alternative venue.
• The business is structurally profitable through a 49% revenue collapse — gross margin never left the 77–82% band through COVID.

Chance of a total loss?

Very low. It would require either the confiscation or invalidation of the VIE structure combined with an inability to access the onshore operating subsidiaries (which hold the great majority of assets), or a forced US delisting combined with an inability to convert into the Hong Kong line. The HKEX dual primary listing is the specific mitigant that distinguishes this from a single-listed ADR. The realistic downside is a permanently lower earnings base at a permanently lower multiple — roughly the mid-US$30s on the bear scenario — not a zero.

Recent News & Events

Has the business environment changed recently?

Yes — decisively, and in a tight, dateable sequence.

INTERPRETATION: two features of this sequence carry analytical weight. First, the guidance cut came twelve days after the rail summons, strongly suggesting the rail action — not the anti-monopoly probe alone — was the proximate trigger for quantifying the Q2 hit. Second, the pricing tool was surrendered voluntarily on 10 March, before any SAMR finding — the Meituan playbook for minimising a penalty, and evidence that management expects an adverse outcome.

Significant acquisitions?

None. The material transaction was a disposal: the June/July 2025 MakeMyTrip sell-down (34,372,221 shares, ~US$3.0 billion, RMB 15.2 billion gain).

Change in accounting policies?

No policy change. But two presentation issues have become materially more important as the numbers grew:

1. The non-GAAP bridge excludes fair-value changes but not realised disposal gains, so FY2025 non-GAAP EPS of RMB 45.59 contains the MakeMyTrip gain. The clean figure (RMB 13.4bn) appears only in verbal Q4-call commentary.
2. The balance-sheet “Investments” line of RMB 54.79 billion includes RMB 22.95 billion of held-to-maturity time deposits and financial products — disclosed parenthetically. The genuine strategic equity portfolio is RMB 31.84 billion. Anyone treating the full line as strategic stakes overstates them by 72%.

Separately, in December 2024 the contractual arrangements with Chengdu Ctrip Travel Agency were terminated and the entity acquired outright as a wholly-owned subsidiary — a modest but genuine reduction in VIE reliance.

Recent changes — new markets, facilities, management?

• New markets: APAC expansion is the strategic priority, with the Trip.com brand pushing into Korea, Japan and Southeast Asia against Agoda. Middle East flagged as an emerging growth region.
• New facilities: first offline flagship store opened in Shanghai (silver-generation segment); one-stop inbound service counters launched at major airports in Beijing, Shanghai and Hong Kong; multilingual self-service facilities at 241 attractions in 16 languages.
• New products: proprietary vertical travel AI models; TripGenie and Trip.Planner; AI-native search integrated with the core search function; third-party agent access via Skills and MCP interfaces.
• Management: stable at the top — James Liang (Executive Chairman, founder), Jane Sun (CEO), Cindy Wang (CFO), Xing Xiong (COO). Board additions: May Yihong Wu joined as a director in February 2026 (concurrently audit-committee chair at MakeMyTrip, and an independent director at Alibaba Health, Swire Properties and Noah Holdings). Ten Form 3s were filed in March 2026, consistent with a cohort of new Section 16 filers being registered.
• Capital structure: US$5.0 billion repurchase authorisation (August 2025); first dividend (February 2025); a US$3.0 billion portfolio monetisation (July 2025).

APPENDIX B — Source Appendix

Report date: 25 July 2026 · All sources accessed 25 July 2026 unless otherwise noted

Sources are ordered by evidentiary priority: regulatory filings first, then earnings releases and transcripts, then regulatory and agency documents, then quantitative data services, then trade and financial press. All sources listed are public and independently verifiable.

1. Primary regulatory filings (SEC EDGAR, CIK 0001269238)

Trip.com Group is a foreign private issuer: it files Form 20-F annually and Form 6-K episodically. It does not file 10-K or 10-Q, and its insiders are exempt from Section 16 reporting under Exchange Act Rule 3a12-3(b). The trailing 60-month corpus (since 25 July 2021) comprises 78 Form 6-K, 5 Form 20-F, 33 Form 144, 10 Schedule 13D/G, 10 Form 3, 1 Form 4, 8 UPLOAD and 5 CORRESP.

All documents above are publicly available from SEC EDGAR.

2. Earnings call transcripts

Sourced via the ROIC.ai MCP transcript tools and read in full.

3. Regulatory and agency sources

4. Quantitative data services

5. Trade and financial press

6. Peer company filings used for comparison

Peer take rates, margins and returns were computed from each company’s own FY2025 reported results.

8. Currency, share count and reference-price conventions

• Reporting currency: Renminbi (RMB). USD conversions use RMB 6.8905 = US$1.00, the company’s own convenience translation rate in the Q1 2026 release (verified: RMB 16,208m → US$2,353m; RMB 3,330m → US$483m).
• FY2025 per-share figures use the company’s own translation (RMB 47.67 = US$6.82, i.e. 6.99), as stated in the Q4 2025 release.
• Reference price: US$43.64, NASDAQ close 24 July 2026 (AZI price history, source #27).
• Share count: 629,705,222 ordinary shares outstanding at 31 March 2026 (Q1 2026 balance sheet). 1 ADS = 1 ordinary share. Diluted weighted average for Q1 2026 was 681,679,206.
• Enterprise value was rebuilt by hand at the 24 July 2026 close rather than taken from any data service, because aggregator EVs for this issuer strike market capitalisation at fiscal-period end and conflate held-to-maturity deposits with strategic equity stakes.