Baidu Q2 2026 Earnings: Watch List After a Missed Defensive Entry

Baidu reported second-quarter results on 18 August, and the headline reads like a company in terminal decline: total revenue of RMB 31.3 billion, down 4% year over year, the fifth straight quarterly decline; online marketing down 19%; diluted earnings per ADS crushed to RMB 5.74 from RMB 20.35 a year earlier; non-GAAP EPS of RMB 7.22 missing consensus by a wide margin. The stock initially sold off to near 52-week lows, dropping to around $90. But under the hood the composition is changing fast: AI Cloud Infra grew 50%, GPU Cloud grew 283%, and Baidu's AI-powered business is now half of General Business revenue, up from 38% a year ago. The market initially priced a dying search engine, but the stock has since rallied well past our target entry zone. Our original plan was to write short puts for a defensive entry, but we missed the window. We are now placing BIDU on the watch list.

At a Glance

• Reported: 18 August 2026; total revenue RMB 31.3B (-4% YoY, -2% QoQ), a fifth straight quarterly decline
• Online marketing: RMB 13.1B, -19% YoY, still 52% of General Business
• AI-powered Business: RMB 12.5B, +25% YoY, now 50% of General Business (was 38%)
• AI Cloud Infra: RMB 7.3B (+50%); GPU Cloud +283% (accelerating from +184% last quarter)
• Earnings: net income RMB 2.3B (-68%); non-GAAP EPS RMB 7.22 vs ~11 consensus (miss)
• Balance sheet: RMB 283.1B (~US$42B) of cash and investments as of 30 June
• Cash flow: operating cash flow +RMB 3.4B, positive for a fourth straight quarter; capex RMB 11.4B; free cash flow negative
• Apollo Go: 28 cities globally, 240M+ fully driverless kilometres; testing with Uber and Lyft in London, live in Dubai, first permits in Hong Kong
• Our move: we initially planned to write at-the-money short puts (90 strike) as a defensive entry. The stock briefly touched our target but has since rallied. We are moving BIDU to the watch list and will reconsider if it drops back near $88.

The Numbers

The shape matters more than any single line. The decline is a mix story: ad revenue fell roughly RMB 3.1 billion year over year while the AI businesses added about RMB 2.5 billion on top. The replacement engine is no longer a side bet; it is half the company, and it is compounding. The costs are real: RMB 11.4 billion of capex in one quarter made free cash flow negative, and a swing in investment fair values and foreign exchange (total other income fell from RMB 4.9 billion to RMB 184 million) did most of the damage to reported profit.

The Valuation

This is where it gets interesting. Baidu holds RMB 283.1 billion (roughly US$42 billion) of cash and investments against a market capitalization of about US$31 billion after the initial sell-off. Net of debt, the balance sheet is worth roughly US$26 billion. In other words, the market was pricing the entire operating business, search, AI Cloud, and the entire robotaxi fleet, at a low-single-digit-billion dollar figure. The market was treating the company as a melting ice cube. Our view is that it is a company in transition, not in terminal decline: the cash funds the transition, the AI stack is now large enough to matter, and the robotaxi business is the option nobody was paying for. However, the recent rally has removed the extreme discount we were looking for.

The Long-Term Thesis: The Cabin

The driverless car interior is the next attention surface. A ride is 15 to 40 minutes of hands-free time, and when the driver disappears, that time becomes monetizable real estate: the operator controls the default screen, the voice assistant, and the ride loop. Baidu owns the whole stack in the market where robotaxis scale fastest: the fleet (Apollo Go, now in 28 cities with 240 million-plus fully driverless kilometres and an outstanding safety record), the assistant (ERNIE), the maps, and China's largest advertising platform to fill the inventory. The global expansion is real: open-road testing with Uber and Lyft in London, fully driverless commercial operations in Dubai, the first fully driverless testing permits in Hong Kong, and partnerships in Switzerland and Kazakhstan. This is a real long-dated option, not a certainty: monetization is years away, and the smartphone in a passenger's pocket remains the strongest competitor. But at the $90 level the option was nearly free, and the market's despair was our entry point.

Our Take

We are not buying shares into a falling knife, nor are we chasing the recent rally. Our original plan was to write at-the-money puts, around the 90 strike, collecting premium for a defensive entry. Unfortunately, we missed the window to execute this trade when the stock briefly dipped to our target level, and it has since rallied significantly higher. We are honest about the state of the transition: the ad decline is still deepening and the AI monetization payoff is not yet visible in the numbers. What made the trade defensible was the price: a ~US$31 billion market cap against roughly US$26 billion of net cash. Since the stock is now much higher, the margin of safety is gone. We are moving BIDU to the watch list. If the stock drops back under $90, closer to $88, we will reconsider initiating a position.

Verdict: Watch list. We missed the short put entry at $90. We will re-evaluate for a defensive entry or long-term hold only if the stock drops back to the $88 range.

Risks

• Ad decline steepens further before AI profits fully replace them, the single biggest near-term risk.
• AI capex pressure: free cash flow is already negative (RMB 8 billion in Q2) and the buildout is not done.
• FX and investment marks: Q2 showed how fast reported profit can be hit by currency and fair-value swings.
• Regulation and geopolitics: Chinese tech policy and the perennial ADR risk.
• Robotaxi monetization is years away; competition from Waymo, Tesla and local rivals is real.