NextFin

Baidu Revenue Falls for Sixth Quarter as AI Cloud Growth Fails to Fully Offset Declines

Summarized by NextFin AI
  • Baidu’s Q3 revenue fell 1% year over year to RMB30.8 billion, marking the sixth consecutive quarterly decline despite renewed growth in General Business.
  • AI Cloud Infra revenue surged 94%, while AI-powered businesses reached 53% of General Business revenue, confirming AI’s growing operational importance.
  • The replacement remains incomplete: AI-native marketing revenue declined 15%, iQIYI revenue fell 10%, and total costs rose faster than consolidated revenue.
  • Baidu’s transition is structurally promising but financially unproven; sustainable success requires AI growth to outpace legacy advertising declines and generate stronger margins.

NextFin News - Baidu’s latest quarter answers one question and leaves the more important one unresolved: AI is now large enough to restore growth in the company’s core business, but not large enough to stop reported consolidated revenue from falling for a sixth straight quarter. Revenue for the three months ended Sept. 30 dropped 1% year over year to RMB30.8 billion ($4.60 billion), while Baidu General Business revenue rose 1% to RMB25.0 billion. The transition is working operationally. It is not yet working at the group level.

The distinction matters because the quarter was not a simple failure of AI monetization. Baidu said AI Cloud Infra revenue increased 94% year over year, and Baidu Core AI-powered Business accounted for 53% of Baidu General Business revenue. Yet online marketing remains under pressure, AI-native marketing revenue fell 15% to RMB2.4 billion, and iQIYI revenue declined 10% to RMB6.0 billion. The result is a business in mid-transition: the new engine is accelerating, but the old monetization model is losing output faster than the new one can replace it.

Management’s own framing captures the tension. “AI Cloud Infra is becoming a new growth engine for Baidu, with an increasing contribution to both revenue and profit,” Chief Financial Officer Haijian He said in the company’s results release. That is a structural claim. The figures show a cyclical and execution problem layered on top of it: Baidu has built demand for AI infrastructure, but it still must convert that demand into durable, profitable revenue while advertising-related businesses shrink.

The Headline Decline Hides a More Important Split

The reported sixth consecutive decline in total sales is the cleanest measure of the pressure facing Baidu, but it is not the cleanest measure of where the operating business is heading. The company’s Q3 revenue of RMB30.8 billion was down from RMB31.3 billion in the second quarter and RMB31.2 billion a year earlier. The quarter’s 1% annual decline was milder than the 4% annual decline reported in Q2, but the improvement did not come from a broad recovery across the portfolio.

Baidu General Business supplied the positive signal. Revenue rose 1% year over year to RMB25.0 billion, reversing the 4% annual decline recorded in Q2. Within that business, AI Cloud Infra grew 94% year over year. The increase was large enough to help offset weakness in online marketing, and management said AI Cloud had become a major contributor to both revenue and profit.

The offset is still incomplete. Baidu reported AI-native marketing revenue of RMB2.4 billion, down 15% from a year earlier. The category is important because it shows that adding AI to the advertising product does not automatically restore advertising economics. Baidu’s AI transformation is creating infrastructure demand faster than it is rebuilding the monetization layer attached to search.

iQIYI added another drag. Its revenue declined 10% year over year to RMB6.0 billion, compared with a 5% annual decline in Q2. Because iQIYI is presented separately from Baidu General Business in the company’s revenue disclosure, its decline obscures the improvement in the core operating segment. The arithmetic explains why a 1% increase in General Business can coexist with a 1% fall in group revenue.

A pre-release analyst estimate put third-quarter revenue at $4.61 billion, or 5.18% annual growth. Baidu reported $4.60 billion, making the dollar result close to that estimate but the growth outcome materially weaker than the estimate implied. That expectation gap does not resolve the strategic question. A company can land near an estimate in dollars and still expose a difficult transition. Baidu did both.

AI Growth Is Real, but the Transmission Mechanism Is Not Yet Complete

The direct story is familiar: AI Cloud grows, legacy advertising weakens, and the two lines partially offset one another. The more useful question is how the growth travels through Baidu’s income statement. AI Cloud first creates usage revenue from computing and infrastructure. It then has to generate enough margin and customer retention to cover the cost of capacity, research, traffic acquisition, and the lost contribution from search advertising. Q3 shows the first step clearly and the second only partially.

AI Cloud Infra’s 94% growth is powerful in isolation, but growth percentage is not the same as economic replacement. In Q2, AI Cloud Infra revenue was RMB7.3 billion, up 50% year over year, after RMB8.8 billion in Q1, up 79%. Those figures show a business that remains large and rapidly expanding, but they also show that revenue can move sequentially with the timing of enterprise deployments and cloud demand. The Q3 release emphasized annual growth rather than providing an absolute AI Cloud Infra revenue figure in the material reviewed here. It is therefore safer to judge the trend by its contribution to General Business than to assume that a 94% growth rate translates directly into cash earnings.

Costs reinforce the point. Baidu’s Q3 cost of revenue rose 9% year over year to RMB19.9 billion, even as total revenue fell 1%. The company attributed the increase primarily to higher cloud-business costs and traffic-acquisition costs, partly offset by lower content costs. The infrastructure engine is scaling through a cost base that is rising faster than the consolidated sales line. That can be rational during capacity build-out, but it raises the hurdle for AI to become a replacement rather than a supplement.

“In Q3, Baidu General Business revenue returned to year-over-year growth, increasing by 1%. AI Cloud Infra revenue grew 94% year over year, helping to partially offset weakness in online marketing and supporting the overall top-line recovery,” Haijian He, Baidu’s chief financial officer, said in the company’s Q3 results release.

The cash-flow data provide a counterweight. Operating cash flow remained positive at RMB2.8 billion, down from RMB3.4 billion in Q2 but still evidence that the group is generating cash from operations while investing in AI. That is different from a strategy dependent entirely on external financing. It does not prove that the new revenue stream earns an attractive return: operating cash flow reflects the consolidated business, including working-capital movements and the legacy franchise.

The transmission mechanism therefore has three stages. First, enterprise customers buy AI computing and services. Second, Baidu must retain those customers and move them toward higher-value applications, rather than merely supplying low-margin capacity. Third, AI applications and AI-native advertising must rebuild the monetization economics that traditional search once provided. The first stage is advancing quickly. The third stage remains the bottleneck.

This Is a Structural Shift With a Cyclical Drag

Baidu’s transition should be classified as structural, but the current sales weakness contains a cyclical component. The structural part is the change in the industry’s revenue architecture: search is no longer the only interface through which users and advertisers reach information, while computing and model services have become strategic products. That shift does not reverse automatically when China’s advertising market improves.

The cyclical part is the pace of monetization. Baidu said its AI-native marketing revenue fell 15% in Q3 because of a challenging macroeconomic environment, changing competitive dynamics, and a strategic decision to prioritize user experience over near-term monetization during its AI Search transformation. Those factors can improve. Advertising budgets can recover, competition can settle, and product changes can produce better conversion rates. But a cyclical recovery in advertising would not restore the old model unchanged, because the product itself is being redesigned.

Historical comparisons from the first three quarters of 2026 show why the two forces must be separated. In Q1, Baidu General Business revenue was RMB26.0 billion, flat sequentially, while AI Cloud Infra grew 79% year over year. In Q2, General Business fell 4% year over year to RMB25.2 billion, and AI Cloud Infra was RMB7.3 billion, up 50% year over year. In Q3, General Business returned to 1% annual growth at RMB25.0 billion, while AI Cloud Infra growth accelerated to 94%. The sequence suggests that AI infrastructure can stabilize the core business even as its growth rate and quarterly mix fluctuate.

But the same history also argues against declaring victory. Baidu App monthly active users fell from 655 million in March to 644 million in June and 604 million in September. User declines do not mechanically determine revenue, and Baidu’s AI products can monetize enterprise demand independently of consumer reach. Still, a shrinking user base makes the advertising replacement task harder: fewer users can mean less behavioral data, less traffic inventory, and less distribution for AI search products.

That is the structural tension. AI Cloud can grow even while the consumer funnel weakens, but the long-term economics will depend on whether Baidu can connect the two businesses. If AI Cloud remains an infrastructure supplier while search loses high-margin advertising, the company may post impressive growth in a lower-quality revenue mix. If AI search drives paid usage and enterprise conversion, the same infrastructure investment can become a platform advantage.

The Second-Order Risk Is Margin, Not Just Revenue

The first-order conclusion is that AI growth is offsetting search weakness. The second-order question is whether the offset is profitable enough to support the investment cycle. Q3’s cost data make that question unavoidable: cost of revenue rose to RMB19.9 billion, up 9% annually, while revenue fell to RMB30.8 billion.

That relationship can produce two very different outcomes. In the favorable case, cloud capacity is installed ahead of demand, so costs rise before utilization. As customer workloads scale, revenue grows faster than the fixed cost base, and margins expand. In the unfavorable case, competition forces Baidu to price infrastructure aggressively, while model development and traffic acquisition remain expensive. Revenue then grows without producing enough incremental profit to compensate for the decline in advertising.

Baidu’s Q2 history illustrates the capital intensity of the favorable case. Capital expenditures nearly tripled year over year to RMB11.4 billion in the quarter, and free cash flow was about negative RMB8 billion in a pre-release analyst preview. Those figures are not Q3 results, but they establish the investment burden that investors were already watching. Q3 operating cash flow of RMB2.8 billion was positive, yet the company’s statement that it will continue investing decisively in AI means the cash-flow question is not finished.

This is where consensus can mislead. The pre-release revenue estimate called for 5.18% year-over-year growth, but the company delivered a 1% decline. The dollar result was close to the $4.61 billion estimate, but the growth profile was not. Analysts were forecasting enough AI and other growth to lift the top line; that lift did not arrive in the reported quarter.

The expectation gap is not necessarily a new bearish thesis. It is a test of what investors are paying for. If the central claim is that AI has become Baidu’s growth engine, then the relevant benchmark is not whether AI revenue is rising. It is whether AI growth is large, profitable, and durable enough to outrun the decline in the legacy business. Q3 passed the first test, partially passed the second, and did not yet pass the third.

Robin Li, Baidu’s co-founder and chief executive, presented the company’s long-term case directly:

“Our long-term investments in full-stack AI are translating into meaningful growth at scale. This quarter, AI Cloud Infra emerged as a major driver of Baidu General Business revenue expansion, and Baidu Core AI-powered Business accounted for 53% of Baidu General Business revenue, further reinforcing AI’s position at the core of our business,” Robin Li said in the Q3 results release.

That statement is credible as a description of mix shift. It is not yet proof of economic replacement. The next proof point must come from the relationship between AI revenue growth, cloud costs, operating income, and cash generation.

The Strongest Counter-Thesis: The Quarter May Mark the Trough

The strongest argument against a cautious reading is that Baidu may already be past the worst point in the transition. General Business returned to annual growth in Q3, AI Cloud Infra growth accelerated to 94%, operating cash flow remained positive, and Baidu Core AI-powered Business accounted for 53% of General Business. On this view, the consolidated decline is a lagging indicator distorted by iQIYI and advertising weakness. The operating business has already turned.

That counter-thesis deserves more than a dismissal. Transitions often look worst immediately before the new product reaches scale. AI infrastructure can have a delayed payoff because customers begin with pilots, then expand workloads after they see productivity gains. Baidu’s full-stack position—from models and chips to cloud and applications—could allow it to retain more of the economics than a pure reseller. If AI Cloud growth remains above 50% and AI-native marketing stops shrinking, General Business could compound even while iQIYI stabilizes.

The company also has evidence that AI is not merely a marketing label. AI Cloud Infra grew in each of the first three quarters, from 79% annual growth in Q1 to 50% in Q2 and 94% in Q3. Core AI-powered Business moved from half of General Business in Q2 to 53% in Q3. Those are operational milestones, not future promises.

But the counter-thesis fails if it treats scale as equivalent to replacement. Total revenue still fell 1%; AI-native marketing fell 15%; iQIYI fell 10%; and App MAUs dropped to 604 million from 655 million in March. The company’s own release said AI Cloud only “partially” offset weakness in online marketing. The replacement engine is running, but the old engine is still losing power.

The falsifying signal for the cautious structural-transition judgment is specific: if Baidu General Business revenue grows at least 5% year over year for two consecutive quarters while AI Cloud Infra growth remains above 50%, AI-native marketing returns to year-over-year growth, and cost of revenue grows more slowly than revenue, the evidence would show that the monetization bottleneck has broken. Conversely, if General Business returns to a year-over-year decline and AI Cloud growth falls below 30% while costs continue to rise faster than sales, the claim that the AI transition is stabilizing the operating business would be wrong.

For now, the data support a middle position. Baidu is not an advertising company waiting for AI to arrive. It is an AI company whose legacy monetization is declining before the new model fully matures.

What the Results Mean Across Time Horizons

In the short term, the most important variable is sentiment around the replacement rate. A sixth total-revenue decline can keep pressure on the stock narrative even if General Business has improved, particularly when the pre-release estimate had looked for growth. No verified current share-price move is included here, so the conclusion rests on the earnings arithmetic: the reported result did not deliver the top-line acceleration that the AI thesis implied.

Over the medium term, investors will focus on operating leverage. The key comparison is not AI Cloud growth against zero; it is AI Cloud growth against cloud costs, capital spending, and the decline in online marketing. Positive operating cash flow of RMB2.8 billion provides time, but it does not remove the need for returns on the AI investment. The next quarters will show whether the 94% cloud-growth figure produces expanding profit contribution or simply a larger, more expensive revenue base.

Over the long term, the opportunity and risk are both structural. Baidu can benefit if its model, chip, cloud, and application stack gives enterprise customers a reason to stay within one ecosystem. It remains exposed if AI search reduces the value of the user traffic and advertising inventory that funded the old ecosystem. The beneficiaries of a successful transition would include AI Cloud and enterprise-service businesses; the exposed assets would be online marketing and any operation whose revenue depends on the old search funnel.

The base case is a slow replacement: General Business remains around low-single-digit growth, AI Cloud expands rapidly but with uneven quarterly revenue, and total Baidu revenue remains constrained by iQIYI and advertising runoff. The upside case requires two triggers—AI-native marketing returns to growth and cloud costs begin to grow more slowly than AI revenue. The downside case is a renewed General Business decline caused by AI Cloud growth falling below 30% or by sustained double-digit contraction in AI-native marketing.

The next decisive data point is not another announcement that AI is strategic. It is the quarterly bridge between AI Cloud growth and consolidated profit. Baidu has demonstrated that AI can become a larger share of the business. It has not yet demonstrated that AI can replace the earnings quality of search advertising.

Baidu’s Q3 was therefore neither a clean turnaround nor a simple AI failure. It was a structural transition advancing through a cyclical and costly monetization gap. The company’s core business has begun to grow again; the group still needs the new engine to run faster than the old one fades.

Explore more exclusive insights at nextfin.ai.

Insights

Why did Baidu revenue fall again?

How fast is Baidu AI Cloud growing?

Can AI Cloud replace lost ad revenue?

Why are AI-native ads declining?

What caused higher cloud costs?

Is Baidu's AI growth profitable yet?

How did Baidu's user base change?

Could AI search weaken advertising?

What does iQIYI's decline mean?

How does Baidu compare with rivals?

Could AI Cloud margins improve soon?

What would prove Baidu's AI pivot works?

What risks could derail Baidu AI?

Will Baidu's core growth continue?

What should investors watch next?

How might Baidu's AI stack evolve?

Why did Baidu miss growth expectations?

Can cash flow fund Baidu AI investment?

Search, ask or analyze anything...