Baidu’s next report lands under conditions no earnings preview has quite addressed. Baidu faces a shareholder class-action lawsuit alleging the company overstated artificial intelligence’s ability to offset declines in its legacy online marketing business. The deadline for lead plaintiff filings is November 13, 2026, three days before the numbers drop on November 16, 2026. That sequencing matters. Q3 is not just a revenue scorecard. It is the first set of financials that lands while that legal claim is active and unresolved.
The underlying tension is real, and the data bears it out. Online marketing services revenue fell 19% year over year in Q2. Against that, GPU Cloud revenue increased 283% year over year in Q2, accelerating from 184% the prior quarter. That is a genuinely extraordinary growth rate. The lawsuit’s contention, though, is not that the AI numbers were fabricated. It claims the company overstated how much its AI-powered business could mitigate the decline in the legacy online marketing business, and that total revenue was reasonably likely to decline regardless.
That framing resets the question investors should ask on November 16. The usual read on Baidu has been: watch AI cloud growth; if it accelerates, the transition thesis holds. The lawsuit introduces a stricter test. Baidu Core AI-powered Business revenue grew 25% year over year in Q2 but declined 8% sequentially. Sequential deceleration inside a year-over-year surge is precisely the kind of detail the complaint focuses on. If Q3 shows another sequential dip in the AI-powered business aggregate, the litigation’s factual foundation gets stronger, not weaker.
One caution: near-term “consensus” figures vary by data vendor, and Baidu did not provide specific guidance for Q3. The company did say after Q2 that its online marketing business had bottomed. Whether that claim holds is exactly what November 16 tests.
Apollo Go’s robotaxi results add a separate variable. Baidu has said Apollo Go is making progress on unit economics, but the company has not provided enough detail publicly to support a clean claim that unit economics had reached breakeven in its largest operating city as of Q2, or that management has guided to stronger margins in overseas markets. What is clear is that Apollo Go’s geographic footprint has expanded over time, and Baidu has disclosed that Apollo Go’s global footprint reached 26 cities as of February 2026.
Investors already confronting disappointing financial results face a roughly 38% year-to-date stock decline as of November 10, 2026. The report on November 16 will not resolve the litigation. But it will determine whether management’s “advertising has bottomed” call from August was right, and whether AI cloud can maintain its trajectory against a tightening cost base. Both answers carry legal weight, not just investment weight.
