Autohome Shipped an AI Butler Into a Market That Keeps Missing the Bottom
The AI agent is now real, the offline network is bigger, and the stock still trades below the cash pile—because the core auto market is shrinking faster than the pivot can scale.
ATHM · Earnings Call · 2026-11-04
The bottom that keeps moving
Autohome's third-quarter call was framed around a familiar comfort: the worst is behind us. Management said “we do expect the domestic auto market remain under pressure in the short term, but we believe it is already at the bottom for the moment.” — Sterling Song, Management - likely CEO or IR Head · 2026-11-04 The problem is that this is the same bottom management has been calling for several quarters. In August, the message was more hopeful: “we believe the market will show kind of a recovery second half of this year.” — Yan Zeng, Chief Financial Officer · 2026-08-20 Three months later, the golden September/silver October season never showed up. “This year in the third quarter, we didn't see the usual strength from the traditional golden September, silver October, strong sales seasons.” — Sterling Song, Management - likely CEO or IR Head · 2026-11-04 Full-year passenger-vehicle retail sales are now projected at 18.5 million units, down 22% year over year, with Q4 down 20%-25%. That is not a bottom call; it is a moving target. This is the central tension of the report: Autohome is executing a strategic pivot while its core market is contracting faster than the pivot can offset. The company's own keyword trail reflects that. The prior-quarter list was heavy with Domestic demand and used car exports; the current list is more operational—offline fulfillment network, service provider, vehicle inspection. That is what a company sounds like when it stops debating the market and starts building around it.
An AI agent finally ships, but not a revenue line
The freshest company-specific development is real: on September 5, Autohome launched Cheese Car Butler, an AI-native application built on its proprietary Cangjie model. It covers browsing, selection, purchase, and ownership, and links online consultation with offline store services. The company's Q3 keyword set puts AI Agent near the top, and the product is the most tangible answer yet to the threat management has been discussing since at least Q1 2026, when AI agents were already framed as a new interaction hub that could disintermediate vertical media platforms. But the admission that matters came in the Q&A.
These investments in technology and products are already reflected in our financial statements, financial results. However, AI not already began to generate revenues in scale. Instead, they serve as the underlying technological capabilities across our various products, helping us optimize the product offerings and improve operational efficiency.
In plain terms: the AI agent is a defensive product and an efficiency tool, not yet a P&L driver. That is not unusual for a launch that is weeks old, but it does mean the market cannot underwrite Autohome as an AI company today. The AI work is necessary to protect the franchise; it is not yet a new growth engine.
The transaction pivot is real—and still tiny
The offline buildout is more substantial than the AI launch, at least in physical footprint. Autohome Mall's online car-purchasing model reached 38 cities with more than 1,700 authorized dealers and over 600 vehicle series. Autohome Good Car had more than 170 franchise stores. Then comes the number that reframes the scale: “over 170 franchise stores, with cumulative transactions of over 700 vehicles.” — Sterling Song, Management - likely CEO or IR Head · 2026-11-04 For a platform with 79.14 million mobile DAUs, 700 cumulative vehicles is a pilot, not a business. The used-car effort is following the same pattern—standardize first, scale later. The full-process platform was extended to six cities, with 265 inspection items, one-on-one service, auction matching, and online transfer. The cross-border export platform averaged more than 3,000 vehicles listed for sale daily and has seen monthly transaction volume rise since its first export order in July. These are sensible building blocks, but they are still sub-scale relative to the hole left by the declining media and lead-generation businesses. Notably, the Haier Group thread that dominated the Q1 2026 call has gone quiet. Back then management said, “After Haier becoming our new controlling shareholder, we don't have material change in our overall strategic direction.” — Sterling Song, Management, likely Investor Relations or Senior Executive · 2026-03-05 The current call mentions Cartech synergies but not Haier as a live narrative. That may be because the story has moved from shareholder structure to store-level execution—but it also means the hoped-for Haier synergy is no longer a headline catalyst.
The cash pile is the valuation
The financials explain why the equity looks so strange. Q3 revenue was CNY1.31 billion, gross margin improved to 67.2% from 63.7%, but operating profit fell to CNY96 million from CNY147 million, and adjusted net income dropped to CNY253 million from CNY407 million. The company remains cash-rich: cash, cash equivalents, short-term investments, and other long-term investments totaled CNY17.65 billion. Against a market cap of roughly $650.9 million, that cash balance is the story. Autohome is returning a large slice of it: the board approved a dividend policy payment of $1.58 per ADS, or about CNY1 billion in aggregate, and the $400 million buyback has already had $155 million executed as of October 30. The buyback alone is enormous relative to the equity value. This is not a growth-stock capital allocation model; it is a cash-return model for a business the market believes is in terminal decline. That is the uncomfortable conclusion. Autohome has shipped a real AI agent, built a broader offline fulfillment network, and kept a fortress balance sheet. But the core market is shrinking, the transaction business is still measured in hundreds of vehicles, and AI is explicitly not yet a revenue line. The stock is not trading on the AI narrative; it is trading on the cash—and on whether management can keep returning it while the pivot scales. Until Good Car and Autohome Mall show throughput, not just store count, Autohome remains a cash-rich, cash-returning, shrinking platform with an AI option.