Summary
The most dangerous mistake when looking at Baidu today is to oversimplify the story as "search advertising is declining, so the company is finished."
But it is also too early to say, "AI will revive Baidu, so buy it."
In the first quarter of 2026, Baidu reported total revenue of RMB 32.075 billion, operating income of RMB 3.193 billion, and non-GAAP operating income of RMB 3.8 billion. Within Baidu General Business, online marketing revenue was RMB 12.6 billion, down 22% year on year. At the same time, Baidu Core AI-powered Business reached RMB 13.6 billion, up 49%, and AI Cloud Infra reached RMB 8.8 billion, up 79%.
In other words, the legacy advertising business is clearly weak. AI is growing.
The problem is how much margin and free cash flow that AI can generate.
According to official disclosure, operating cash flow in 2026 Q1 was positive at RMB 2.670 billion, but FCF after capex was negative RMB 3.246 billion on a Baidu Inc. basis. As long as Baidu keeps investing in AI cloud, inference infrastructure, and robotaxis, revenue growth alone will not be enough for stock revaluation.
This article organizes Baidu's AI transition, search advertising decline, AI cloud margins, Apollo Go profitability, and 2027 turnaround scenario as of May 31, 2026.
The Market No Longer Views Baidu Only as a Search Company
Baidu used to be easy to understand.
It was China's largest search company and earned high margins from search advertising. Investors valued it through that lens.
But in 2026, that simple view is no longer enough.
The capital market is weighing two forces: the decline risk of a search advertising company and the revaluation potential of a Chinese AI infrastructure company.
In 2026 Q1, Baidu General Business revenue was RMB 26.0 billion, up 2% year on year and flat quarter on quarter. Inside that, online marketing services revenue was RMB 12.6 billion, down 22% year on year. Its share of General Business also declined to 48%.
On the other hand, other revenue was RMB 13.4 billion, up 42% year on year, mainly driven by AI cloud growth.
The structure is clear.
As a search advertising company, Baidu is under pressure. As an AI infrastructure and AI application company, it is growing.
The market wants to know whether the latter can truly replace the profit of the former.
The Search Dilemma: Better AI Can Weaken the Ad Model
Baidu faces a dilemma similar to Google's.
Adding generative AI to search improves user experience. Users can reach answers through AI summaries or responses without clicking through multiple links.
But from a search advertising perspective, this is not simple.
Traditional search advertising was built on search result pages, clicks, keywords, and ad slots. The more AI summarizes answers, the narrower the old click path becomes.
That is Baidu's real difficulty.
If it slows AI search, it risks losing on user experience. If it pushes AI search forward, it can damage existing ad revenue.
That is why the market is watching AI-native Marketing Services more than model performance alone.
AI-native Marketing Services revenue was RMB 2.3 billion in 2026 Q1, up 36% year on year. That is not bad, but it is still small compared with the pressure from RMB 12.6 billion in online marketing revenue.
The headline "AI advertising is growing" is not enough. The real issue is how quickly and at what margin AI advertising can offset the decline in traditional advertising.
The Blind Spot in 49% AI Growth: Investors Care About Quality
The most eye-catching figure in Baidu's 2026 Q1 results was Baidu Core AI-powered Business revenue of RMB 13.6 billion, up 49% year on year.
It also accounted for 52% of Baidu General Business. Looking only at that number, Baidu already appears to be shifting into an AI company.
But investors are unlikely to respond to revenue growth alone.
The components need to be separated.
| AI-Related Item | 2026 Q1 | YoY Change | Interpretation |
|---|---|---|---|
| Baidu Core AI-powered Business | RMB 13.6 billion | +49% | Core of the AI transition. 52% of General Business |
| AI Cloud Infra | RMB 8.8 billion | +79% | Main growth driver. Strong demand but heavy capex burden |
| AI Applications | RMB 2.5 billion | Flat | Application monetization remains weak |
| AI-native Marketing Services | RMB 2.3 billion | +36% | Candidate to replace search ads, but still limited in scale |
The key point is that not all AI growth has the same quality.
AI Cloud Infra is growing, but GPUs, data centers, networks, power, and depreciation are heavy. AI Applications were flat and do not yet look like explosive monetization. AI-native Marketing Services are growing, but not yet large enough to quickly fill the hole in legacy advertising.
The numbers are strong. But this is also why the market is not fully convinced.
Can AI Cloud Restore Margins?
Baidu's AI cloud looks attractive if judged by growth rate alone.
AI Cloud Infra reached RMB 8.8 billion in 2026 Q1, up 79% year on year. GPU Cloud was reportedly up 184% year on year.
But AI cloud is a very different business from search advertising.
Once a search advertising platform is established, it can generate high margins. AI cloud is much more affected by capex, inference cost, GPU utilization, and price competition.
In China's cloud and AI market, Alibaba, Tencent, Huawei, and ByteDance-linked services are also competitors. Even if customer demand grows, price competition can still compress margins.
The main checkpoints are these:
| Checkpoint | Why Investors Care |
|---|---|
| Inference cost | As AI search and agents are used more, compute cost per use affects margins |
| Infrastructure utilization | Low data-center or GPU utilization increases depreciation burden |
| In-house chips and supply chain | Kunlun-related chips may help control cost and supply constraints |
For Baidu to be revalued as an AI company, AI cloud revenue growth alone is not enough.
Non-GAAP operating margin, adjusted EBITDA margin, and FCF need to improve in the same direction.
In 2026 Q1, non-GAAP operating margin was 12%, and adjusted EBITDA margin was 19%. These are not poor numbers, but the question is how sustainably they can improve while AI investment continues.
FCF Has Not Turned Yet
The most important lens in this draft is margin and free cash flow.
AI companies can look strong if investors focus only on revenue growth.
But the final thing investors want to see is cash.
Official disclosure showed operating cash flow of RMB 2.670 billion in 2026 Q1. Looking only at this, the core business still has cash generation.
But after capex, FCF was negative RMB 3.246 billion on a Baidu Inc. basis.
| Item | 2026 Q1 |
|---|---|
| Operating cash flow | +RMB 2.670 billion |
| Capital expenditures | -RMB 5.916 billion |
| Free cash flow | -RMB 3.246 billion |
This number gives a sober view of Baidu's AI transition.
AI businesses are growing. Operating cash flow is positive. But FCF has not yet absorbed the investment burden.
Therefore, the most important part of the 2027 turnaround scenario is not revenue.
Will AI investment peak? Will inference costs fall? Will AI cloud margins rise? Will Apollo Go's loss burden ease?
Unless these appear in the numbers, the market will continue valuing both the AI growth story and the heavy investment burden at the same time.
Can Apollo Go Become a Second Search Business?
Apollo Go is the biggest dream option in Baidu's portfolio.
In 2026 Q1, Apollo Go provided 3.2 million fully driverless rides, and its weekly peak in March exceeded 350,000 rides. Cumulative rides exceeded 22 million as of April 2026.
This is well beyond the early demonstration stage.
But as an investment, ride count alone is not enough.
Robotaxis involve vehicle costs, sensors, maintenance, remote monitoring, insurance, accident response, municipal approvals, and different regulations by city. Overseas expansion adds geopolitics, data rules, and local transport systems.
Investors should watch four points.
| Issue | What to Watch |
|---|---|
| Ride volume | Whether rides keep growing by city |
| Unit economics | Whether profitability improves per vehicle and per ride |
| Monitoring and operations cost | How much full driverless operation can reduce labor cost |
| Regulation and insurance | Whether commercial expansion is supported by the policy environment |
Apollo Go could become a second search business.
But it still needs proof before it can be treated as a high-margin cash cow like search advertising. For now, it is more natural to view it as a large option value, not yet a business that steadily lifts FCF.
Numbers to Watch in the 2027 Turnaround Scenario
Baidu's 2027 turnaround scenario is clear.
Even if search advertising continues to decline, AI Cloud Infra, AI-native Marketing Services, and Apollo Go grow, while margins and FCF improve.
That is the market's desired shape.
But expectations alone cannot sustain stock revaluation.
The numbers to watch are these:
| Checkpoint | Desired Change in the Turnaround Scenario |
|---|---|
| Online marketing revenue | Decline rate slows |
| AI-native Marketing Services | Moves closer to a scale that can offset legacy ads |
| AI Cloud Infra | Maintains high growth while improving margins |
| Non-GAAP operating margin | Recovers from 12% |
| FCF | Turns from negative to stable positive |
| Apollo Go | City-level profitability, overseas expansion, and lower operating costs become visible |
FCF is especially important.
AI investment tends to show up first as cost, not revenue. GPUs, data centers, model development, talent, and robotaxi vehicles all require capital.
If that investment is truly creating value, it should eventually come back through operating cash flow and FCF.
If it does not, AI may remain a growth story while leaving little cash for shareholders.
Conclusion: Baidu's Revaluation Depends on Cash, Not AI Revenue Alone
Baidu's core problem is not simply the decline in search advertising.
The market is watching when and at what margin AI cloud, AI-native Marketing Services, and Apollo Go can replace the lost advertising profit.
2026 is a transition year.
Online marketing is weak. AI is growing. Operating cash flow is positive, but FCF is still negative. That is Baidu's current position.
If Baidu truly turns around in 2027, the key evidence will not be flashy AI announcements.
It will be non-GAAP operating margin, AI cloud profitability, Apollo Go unit economics, and FCF improvement.
Baidu's future will not be decided only by model performance.
It will be decided by how much stable margin and free cash flow the company can generate through AI.
This article is intended to organize investment thinking and does not recommend buying or selling any specific security. Chinese stocks, Hong Kong stocks, and U.S.-listed ADRs involve price fluctuation risk, currency risk, liquidity risk, regulatory risk, geopolitical risk, and risks related to differences in accounting and disclosure systems.
Sources
- Baidu, Inc., "Baidu Announces First Quarter 2026 Results" https://ir.baidu.com/news-releases/news-release-details/baidu-announces-first-quarter-2026-results/
- Baidu, Inc., "Form 20-F for the fiscal year ended December 31, 2025" https://ir.baidu.com/sec-filings/sec-filing/20-f/0001193125-26-109289/