Baidu Q3: Online Marketing Revenue Falls 22% as Robin Li Puts AI Search Experience Ahead of Monetisation

Baidu has put a number on what its “AI Search transformation” is costing. In its third-quarter 2026 results, published on 16 November, online marketing revenue fell 22% year over year to RMB 11.9 billion. That is the lowest figure in the table Baidu gives, and for the first time advertising makes up less than half (48%) of Baidu General Business revenue.

Meanwhile AI Cloud Infra rose 94% to RMB 8.2 billion. That is what kept General Business growing, by 1% to RMB 25.0 billion. Group revenue was RMB 30.8 billion, down 1%, which Reuters notes was almost exactly in line with consensus.

The search numbers that matter

  • AI-native marketing services: RMB 2.4 billion, down 15% year over year and 7% on the quarter. This is the line meant to hold the new ad formats inside AI search, and it is shrinking.
  • Legacy Business (classic search and feed ads, as Baidu now labels them): RMB 9.4 billion, down 24% year over year and 10% on the quarter.
  • Baidu App MAU: 604 million in September, according to the earnings call.
  • ERNIE Assistant: daily average conversation rounds up 126% year over year in September.

Robin Li’s explanation on the call was blunt. Baidu “will continue to prioritize user experience over near-term monetization as we advance our AI search transformation”, and that “will weigh on our online marketing revenue and overall profitability in the near term.” CFO Haijian He added that advertising “will remain under considerable pressure from macro conditions, intense competition and our ongoing AI search transformation.”

What Baidu didn’t say

As PPC Land points out, there is nothing about what has actually changed on the results page. Baidu gives no share of queries that trigger an AI answer, no ad load inside those answers, no click-through data and no query volume. The only hard evidence of the transformation is financial, and it says that both the old ads and the new ones are earning less.

Baidu did describe the product direction. AI answers are “more organized, more concise and easier to navigate”, and AI search is being tied more closely to ERNIE Assistant, which asks follow-up questions and takes agentic actions. Li also again pointed to ERNIE Light Deep Thinking as “a model we built especially for AI search”. That’s the model Baidu turned into a public API at Baidu World last week.

My take

For anyone doing SEO or paid search in China, this is the clearest admission yet that Baidu is happy to trade clicks for conversations. It fits the QuestMobile data showing search apps losing almost a billion hours in September while AIGC apps gained about the same. Users are moving into chat interfaces, and Baidu is following them into ERNIE Assistant rather than defending the ten blue links.

A few practical points:

  • Expect fewer, more expensive ad slots. A 22% revenue fall with an ad product that is supposed to be “AI-native” means either lower ad load or weak pricing in AI answers. Advertisers on Baidu Phoenix Nest should watch CPCs and impression share closely this quarter.
  • Being cited inside the AI answer now matters more than ranking. If answers are getting “more concise”, fewer sources get shown. Structured, authoritative pages on Baidu-owned properties (Baike, Zhidao, Baijiahao) and well-structured brand sites remain the safest bet. The brand-signature rules for search widgets are a sign of where Baidu wants brands to sit.
  • Don’t expect transparency. Baidu reports a 126% rise in ERNIE Assistant conversation rounds but no referral or click data. Your own analytics are the only measure of what this shift is doing to Baidu organic traffic.

Baidu is now a cloud company that also runs a search engine, and it said so on the call. The search engine is being rebuilt around answers, and for now the advertising side is paying for that.

Dan Taylor is the founder and editor of International Search Insights (ISI), the publication behind ReadIsi, and a Partner and Head of Innovation (Search & AI) at SALT.agency. He has worked in search for more than a decade, in-house and agency-side, and is regarded as one of the West’s leading voices on search engines beyond Google, including Yandex, Baidu, Naver and Cốc Cốc. On ReadIsi, Dan covers the engines and AI answer platforms that English-language coverage tends to miss: Baidu and ERNIE, Yandex and Alice AI, Naver, Cốc Cốc, Shenma, Sogou, WeChat and Weibo search, plus DeepSeek and Alibaba’s Qwen. He also keeps dated records of documented changes, such as the Yandex, Baidu and Naver algorithm update histories and the DeepSeek, ERNIE and Qwen model update histories. His approach is to take learnings from global search engines and apply logic with a high evidence threshold. A claim needs a source, patterns seen in one market are treated as a hypothesis until they are checked in another, and where the evidence is thin he says so. For brands, that means advice built on what can be shown, with measurement that separates being cited from being chosen. Dan is credited with coining the term “edge SEO” and won the inaugural TechSEO Boost research competition in 2018 for his work on CDN and edge changes for SEO. In 2026 he was credited with coining “Infinite Tail” to describe how keyword research is changing with AI search. He is the author of AI Search Marketing For The Travel Sector, writes for Search Engine Journal, Search Engine Land and Travel Marketing Guy, and has advised technology, SaaS, travel and enterprise ecommerce brands, including Cloudflare, GitLab and Airbnb. He has given multiple talks and panels at the International Search Summit in Barcelona, and has also spoken at Optimization Moscow, brightonSEO, Google Search Central Live Deep Dive Asia in Bangkok, Athens SEO and the Asia SEO Virtual Summit 2026, where his talk looked at how Baidu, Naver, DeepSeek, Tencent, Cốc Cốc and Yandex built their own AI answers.