Technology WWC APAC Desk

Baidu AI Monetization Faces a Reality Check as Cloud Revenue Jumps 94%—But Can Profits Keep Up?

Baidu AI Monetization Faces a Reality Check as Cloud Revenue Jumps 94%—But Can Profits Keep Up?

BEIJING, China — Baidu reported a 94% year-on-year increase in AI Cloud Infra revenue for the third quarter of 2026 in financial results released on Nov. 16. The results put Baidu AI monetization under renewed scrutiny because the company’s expanding AI operations are growing alongside a sharp decline in its established online advertising business.

The central issue is no longer simply whether customers will use artificial intelligence. It is whether revenue from those services can support the costs of computing infrastructure, product development and the transition away from traditional search advertising.

Baidu AI monetization: The numbers behind the transition

Baidu’s latest financial disclosure put third-quarter AI Cloud Infra revenue at RMB8.2 billion, compared with RMB4.2 billion a year earlier. Revenue from AI Applications was RMB2.6 billion, essentially unchanged year on year, while AI-native Marketing Services revenue declined 15% to RMB2.4 billion.

Taken together, the three activities generated RMB13.2 billion in what Baidu calls its Core AI-powered Business. That was 37% higher than a year earlier and represented 53% of Baidu General Business revenue, compared with 39% a year earlier.

The proportions matter because they measure a change in what generates sales, not merely how frequently management discusses AI. Yet the categories are management-defined operating measures derived from internal accounts, which the company says have not been audited separately. They should not be mistaken for standalone audited profit figures.

The company’s consolidated revenue was RMB30.8 billion, down 1% year on year. A strong AI subcategory therefore does not mean the entire company is experiencing equally strong growth.

Cloud demand is growing faster than applications

The most striking expansion came from AI infrastructure. Within AI Cloud Infra, GPU Cloud revenue increased 191% year on year. Baidu also reported a 1,159% annual increase in revenue from external customers’ token usage on its Qianfan model-as-a-service platform.

These growth rates suggest rising commercial use of computing and model services. They do not disclose the size of each business’s profit contribution, and a large percentage increase can reflect a comparatively small starting base.

Baidu described stronger demand, higher customer prepayments and a shift towards longer-term contracts in its cloud business. Those developments may improve revenue visibility, although the costs of serving contracts and maintaining capacity remain important to long-term returns.

For readers following China AI investment concentration, this creates a useful distinction: expansion supported by paying customers is different from an investment case resting mainly on expectations of future adoption.

Advertising weakness complicates the growth story

Baidu’s established online marketing services generated RMB11.9 billion during the quarter, down 22% from RMB15.3 billion a year earlier. Online marketing consequently accounted for 48% of Baidu General Business revenue, compared with 62% in the corresponding quarter of 2025.

AI-native Marketing Services also fell, and the company explained that it had chosen to prioritise user experience over near-term monetization during its AI Search transformation. That approach may help build usage, but it can delay the conversion of new experiences into advertising income.

Analysis: Baidu faces two commercial challenges at once. It must keep developing AI search products that people want to use while building a replacement revenue mix strong enough to offset pressure on conventional advertising.

The challenge is not necessarily evidence of a failing strategy. It is evidence that changing the technology powering a service and changing how that service earns money are separate tasks.

Profits and cash flow require a separate test

Baidu recorded RMB626 million in operating income and RMB1.4 billion in net income attributable to the company for the third quarter. Research and development expenses totalled RMB4.8 billion, while operating cash flow was RMB2.8 billion.

These figures demonstrate that group revenue, earnings and cash generation can move differently. They also show why revenue growth alone cannot determine whether a particular AI initiative generates an attractive financial return.

The company reported RMB276.6 billion in total cash and investments as of Sept. 30, 2026. Financial resources provide flexibility, but they are not evidence that every new computing project will earn back its investment.

That is also the central question in the wider discussion of AI infrastructure financing risks: customers may be buying AI services today, but suppliers must still cover their long-term costs.

New products create additional routes to paying customers

Baidu’s November product announcements included ERNIE 5.5 Flash Preview and Token Exchange, a marketplace intended to connect AI model service providers with customers. The company also highlighted its DuMate work agent and Famou Agent 3.0 for enterprise use cases.

Such products can broaden how businesses consume and pay for AI. However, the quarter’s disclosure does not provide separate profitability figures for each new product, nor does it establish that all of them have already achieved sustained commercial returns.

Analysis: Investors evaluating this strategy should distinguish product availability, adoption and profitability. A launch demonstrates that a service exists; usage shows engagement; recurring revenue and sustainable margins provide stronger evidence of economic value.

The same discipline applies to concerns about circular financing in the AI industry. Commercial activity is more informative when readers can understand who ultimately pays, what they receive and whether the underlying service makes economic sense.

What happens next

Baidu’s subsequent financial disclosures will show whether cloud growth remains strong, whether AI Applications revenue accelerates and whether online marketing stabilises. Operating profit, cash generation and any additional detail about product economics will be equally important.

For now, the official figures support a measured conclusion. Baidu is converting significant AI demand into revenue, particularly in infrastructure. They do not establish that the entire AI portfolio has achieved durable profitability, or that an industry-wide financial bubble exists.

The test is whether the new business mix becomes self-sustaining as the company continues funding its technology transition.

FAQ

How fast is Baidu’s AI cloud business growing?

Baidu reported RMB8.2 billion in AI Cloud Infra revenue for the third quarter of 2026, up 94% year on year.

Is Baidu making money from AI?

Its filing reports revenue from AI infrastructure, applications and AI-native marketing. It does not separately establish the net profitability of every AI product or business category.

Why is Baidu’s advertising revenue falling?

Baidu reported a 22% annual decline in online marketing services revenue and said it was prioritising user experience over near-term monetization as it transforms AI Search. The filing does not attribute the entire decline to that single decision.

Does Baidu’s growth prove there is no AI bubble?

No. Customer revenue is evidence of demand, but determining whether investments are economically sustainable also requires understanding margins, capital costs and future cash flows.

What is Baidu’s Token Exchange?

It is a marketplace introduced at Baidu World 2026 to connect AI model service providers with customers. Its individual financial contribution was not separately disclosed in the quarterly release.

Get our stories first on Google

More in Business

See all in Business