Alibaba Group reported a sharp increase in its artificial intelligence cloud business for the quarter ended September 30, even as new US export controls threaten to complicate the company’s access to advanced computing hardware. Revenue from AI Cloud and Compute Services rose 53% year over year to 61 billion yuan, or about $8.6 billion, the company says’s September-quarter results released on November 19. Revenue from external cloud customers grew faster, with recurring revenue from those customers increasing 67%. The results show that demand for AI-related infrastructure is becoming a material growth driver for Alibaba’s cloud division.
Group revenue reached 271.5 billion yuan, up 10% from a year earlier. Alibaba said the acceleration in cloud revenue was supported by adoption of AI-related products and improved profitability. Revenue from AI-related products recorded triple-digit growth for the thirteenth consecutive quarter, the company says’s statement. The company also reported a substantial improvement in operating income and adjusted EBITA, although those measures include the performance of businesses beyond cloud computing. The results So show a fast-growing AI business within a larger group, rather than a standalone disclosure of AI profitability.
Cloud growth is accelerating from external customers
The distinction between internal and external demand matters for a cloud provider. Internal consumption can help a company build and test infrastructure, but external recurring revenue is a clearer indicator that customers are paying for services on an ongoing basis. Alibaba said recurring revenue from external cloud customers grew 67% year over year, excluding compute-equipment solutions revenue. That figure suggests that AI services are being adopted beyond one-off hardware transactions. It does not, by itself, show how much of the growth came from training, inference, model APIs or other cloud products, because the company’s public summary groups those activities within a broader business category.
Chief Executive Eddie Wu said Alibaba’s full-stack AI capabilities had improved commercialisation and showed recent upgrades spanning proprietary chips, supernode architecture, agent-native cloud infrastructure, AI models and agents. The company also identified its new Zhenwu V900 AI chip. Building an in-house silicon portfolio can give a cloud provider greater control over hardware supply and system design, but it does not automatically remove dependence on external manufacturing capacity, advanced packaging, memory and other components. The performance and availability of a chip depend on the wider production ecosystem.
The company’s outlook is complicated by the US Commerce Department’s new rules on advanced semiconductors and remote access to computing capacity. Alibaba’s management said the restrictions would affect products and services planned for the next year, while expressing confidence that it could increase production capacity for its T-Head chips by 2027. Reuters Breakingviews reported that Wu discussed a potential cloud-revenue growth target of 30% to 45% in 2027, with a possible return to 50% or higher in 2028. Those are management expectations, not guaranteed outcomes.
Domestic silicon becomes a strategic requirement
The new controls reach beyond direct shipments of advanced chips. They also place requirements on certain overseas data centres and manufacturing providers that serve customers headquartered in restricted jurisdictions. That matters to Chinese cloud companies because access to compute can be obtained through rented capacity and owned hardware. If a provider cannot rely on the same overseas supply routes, it may need to increase the use of domestic accelerators, optimise workloads for available chips and manage a more fragmented infrastructure base.
Alibaba’s strategy So has two linked parts: expand cloud services that customers are already buying and build enough hardware and software capability to support that expansion under tighter supply conditions. Its own chip development may help with cost, product integration and supply planning, but the company still needs manufacturing partners and a dependable component ecosystem. The results announcement does not disclose the number of Zhenwu chips that can be produced, their performance relative to competing accelerators or the share of workloads that can migrate to them.
Alibaba’s results also need to be read alongside the group’s wider operating mix. Customer-management revenue declined 5% year over year, while the company said the figure would have grown 1% on a like-for-like basis excluding the impact of a business-development programme. The company’s operating income increased sharply, but it attributed much of the improvement to better operating results in quick commerce and cloud. That context matters because the headline growth in AI cloud is occurring within a group that is still managing changes in e-commerce, local services and other businesses. Investors should not treat the cloud segment’s growth rate as a proxy for the performance of the entire company.
The company’s next reports will need to show whether AI cloud growth converts into sustained margins while infrastructure spending remains high. It will also need to explain how export controls affect customer demand, hardware procurement and the timing of new services. For now, the September quarter shows strong demand and better cloud monetisation, while the regulatory environment adds a new constraint to the supply side. Alibaba is growing its AI cloud business at a rapid pace, but its ability to maintain that growth will depend partly on whether domestic silicon can scale quickly enough to offset restrictions on advanced foreign chips.