History

Figures converted from CNY at historical FX rates — see data/company.json.fx_rates for the rate table. Ratios, margins, and multiples are unitless and unchanged.

The Story, As Management Told It

Autohome spent five years rebranding the same story while the numbers moved against it. The 2021 identity — a high-margin "4.0 ABC + SaaS" data platform for automakers — survives in the boilerplate, but operational reality has migrated to physical retail in lower-tier cities, AI products bolted onto a shrinking media business, and a controlling-shareholder swap from Ping An to Haier. Capital returns are real and growing; the growth story has been re-cast roughly every 18 months and the revenue line is lower today ($922M in FY25) than at the start of the period ($1,139M in FY21). Management's narrative credibility has eroded; their cash-return credibility has improved.

1. The Narrative Arc

The cleanest way to read this five-year arc is as four chapters layered on a single trajectory of slow erosion. Each chapter swapped a slogan; none arrested the slide.

2. What Management Emphasized — and Then Stopped Emphasizing

Four patterns stand out. Quiet replacement of partners and product names — the Ping An partnership was a centerpiece for three years, then disappeared from the FY25 business overview entirely; the FY22 "CRM Plus" framework, the FY23 ERNIE Bot partnership, the FY24 DeepSeek mention, and the FY21 catalog of data-product brands (SiNan, Smart QC, EV Smart Cloud) were each promoted as the next thing and then silently swapped for new branding. The slow rotation of the growth engine — the dominant strategic theme moved from "data products for automakers" (FY21-22) to "offline retail in lower-tier cities" (FY24) to "AI core engine + Autohome Mall transactions" (FY25). One never-dropped theme — NEV brand revenue growth has headlined every single call, with the absolute size of NEV revenue never disclosed. One genuinely consistent commitment — share buybacks and the ~$214M annual dividend started small in FY21 and intensified each year, the only commitment that survived four strategy resets unaltered.

3. Risk Evolution

The headline US-listing risks (HFCAA, CSRC) panicked in FY21, softened sharply after the December 2022 PCAOB vacatur and the March 2023 CSRC Trial Measures, and by FY25 read as administrative plumbing. In their place rose risks that are harder for management to control: NEV automakers selling direct and bypassing the ad platform, a nationwide price war that has more than half of Chinese dealers operating at a loss, and a new AI-reliance risk added as its own top-summary headline in FY25. The Ping An "substantial influence" risk was simply renamed "Haier substantial influence" in FY25 — same template, new controller. The Investment Company Act risk introduced in FY23 onward is a quiet acknowledgement that the ~$3.0B cash pile is becoming disproportionate to the operating business it sits on top of.

4. How They Handled Bad News

Two patterns dominate. The first instinct is always to blame an exogenous shock — chips, COVID, raw materials — even when the actual driver (NEV direct sales) was visible by 2021. The honest reframing tends to arrive only when it is unavoidable — the ICE-vs-NEV revenue split was not articulated until FY24, three years after the FY21 collapse. The lone exception is the Q3 FY25 gross-margin reset, where management told analysts directly that transaction-business gross margin "cannot be compared" with the legacy media business — a pre-emptive admission rather than a walk-back, which is the right tone but four years late.

5. Guidance Track Record

Credibility score: 5/10. Two facts drive the number. Capital-return commitments — three successive $200M buyback authorizations and the ~$214M annual dividend floor — have been fully and visibly honored. Operational and growth commitments have a near-perfect missed record: the NEV-retail revenue promise, the Satellite Plan store count, the recurring "H2 recovery" call, the "margins will not be affected" assertion, the TTP asset-light breakeven path. A reasonable investor should fully trust the cash-return mechanics and heavily discount every fresh growth claim until it shows up in absolute reported numbers rather than YoY growth rates against shrinking bases. The fact that no CEO appeared on the Q4 FY25 call — and that the company is now on its third CEO in roughly 18 months — is an additional governance discount factor.

6. What the Story Is Now

The current story has been deliberately simplified. Autohome is no longer pitching itself as a software-and-data platform on top of an advertising business. The FY25 20-F describes it as "a comprehensive automotive service ecosystem" with AI as the core engine and Vehicle Sales as a new top-line revenue category. The legacy media business is shrinking and being managed as a cash source; the growth bet is on Autohome Mall transactions, the Satellite Plan store network in lower-tier cities, a proprietary "Cangjie" LLM with a "Tianshu" service platform, and an international site covering 1,900+ models from 52 Chinese brands across six countries.

What has been de-risked. The HFCAA and CSRC overhangs that dominated FY21 disclosure are effectively resolved. The ownership transition from Ping An to Haier closed cleanly at a $1.8B cash transaction — an arm's-length validation of equity value at a strategic-buyer level. The capital-return policy is durable, predictable, and growing, supported by a ~$3.0B cash and short-term investment pile that is the largest comfort in the story. Disclosure on industry pain (price war duration, 50%+ dealer losses, 70%+ used-car loss makers, average used-car prices -12% YoY) is candid and quantified rather than evasive.

What still looks stretched. Revenue has now declined in four of the last five years. Gross margin has compressed from ~85% in FY21 to 63.7% in Q3 FY25 — a level management says will not recover because the transaction-business mix is structurally lower-margin. Operating cash flow has fallen from ~$535M to ~$129M over the period, while capital returns rose from ~$76M to ~$357M annually — sustainable only as long as the cash pile lasts (which is years, but not indefinite). Engineer headcount is down 18% from peak while AI is being positioned as the "core engine," a tension worth watching. The 500-satellite-store-by-end-2025 target was quietly replaced with a vaguer 1,000-store three-year goal. And the leadership instability — three CEOs in 18 months, no CEO present on the Q4 FY25 call — is a real signal that the post-Haier governance model is still settling.

What to believe and what to discount. Believe the cash-return mechanics, believe the disclosed industry conditions, believe that Ping An risk has genuinely been replaced by Haier risk rather than removed. Discount the NEV growth-rate headlines until absolute NEV revenue is disclosed; discount every new product-line acronym launched since 2021 (data products, ERNIE, DeepSeek, Cangjie, the five "AI Brain/Master/Champion/Acquisition/Inspector" suite) until it shows up as a reported revenue category; discount the assertion that this strategy reset is different from the previous three resets until two more years of execution make the "service ecosystem + AI core engine" identity look like a destination rather than a slogan.