Alibaba's AI Cloud Enters Profit Release Phase: AI Business Profits Double, Margin Rises to 12%, Self-Developed Chips Serve 650+ Clients

Alibaba's AI business achieved both revenue and profit growth in the previous fiscal quarter: Alibaba Cloud's external commercial revenue increased by 45% year-over-year, setting a new high for growth rate in 22 quarters, while the profit margin of its AI cloud and computing power businesses simultaneously rose to 12%. This progress was driven by the dual advancement of scaled commercial use of self-developed chips and accelerated data center delivery. These results indicate that AI investment is shifting from a phase of scale expansion to one of profit release.

Financial reports show that in the company's first fiscal quarter ended June 2026, revenue from Alibaba Cloud's AI-related products reached RMB 12.376 billion, maintaining triple-digit year-over-year growth for the 12th consecutive quarter. Metrics measuring profitability revealed that Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) for the AI cloud and computing services business surged 133% year-over-year, with the adjusted EBITDA Margin rising to 12%.

During the earnings conference call, Alibaba's management further disclosed that the Annual Recurring Revenue (ARR) for AI-related products exceeded RMB 49.5 billion, accounting for 35% of Alibaba Cloud's external commercial revenue.

The key driver behind the margin increase is that the gross margin of AI-related products is significantly higher than the average for cloud products. This means that the high growth in AI revenue is beginning to translate into incremental profits, rather than just "revenue growth without profit growth." Alibaba Group CEO Eddie Wu stated that AI commercialization has moved from the "crossing the inflection point" phase in the fourth fiscal quarter to a stage of "accelerated growth and rising margins" in the first fiscal quarter.

The foundation for this profit release lies in the leap forward in self-developed chips and infrastructure efficiency. T-Head's self-developed Zhenwu M890 chip now serves over 650 external clients, with super-node instances launched on Alibaba Cloud running inference for large models with over 2 trillion parameters. The delivery cycle for large-scale AI data centers has been compressed to 100 days, and management expects AI computing assets to break even in approximately three years.

Synchronous Rise in Revenue and Profit

Alibaba Cloud's external commercial revenue grew 45% year-over-year this quarter, marking the highest growth rate in 22 quarters. AI-related product revenue amounted to RMB 12.376 billion, achieving triple-digit year-over-year growth for the 12th consecutive quarter, serving as the core engine accelerating cloud business growth.

More noteworthy is the simultaneous improvement in earnings quality. EBITDA for the AI cloud and computing power segment surged 133% year-over-year, with the adjusted EBITDA Margin rising to 12%. The Annual Recurring Revenue (ARR) for AI-related products exceeded RMB 49.5 billion, increasing its share of Alibaba Cloud's external commercial revenue to 35%, with gross margins significantly higher than the cloud product average.

Eddie Wu stated that an increasing number of customers are adopting full-stack AI products covering AI agents, models, cloud infrastructure, and self-developed chips. Coupled with expanded business scale and improved pricing power due to tight computing supply, AI is evolving from merely a revenue growth engine for the cloud business into a significant driver of profit improvement. Whether this profit inflection point is firmly established remains to be verified by data in subsequent quarters, but the substantial improvement in quarterly profitability has already signaled its emergence.

Scaled Commercial Use of Self-Developed Chips

The underlying support for profit release is the commercial implementation of T-Head's self-developed chips. T-Head has built a full-stack self-developed system covering GPUs, CPUs, and network chips. Its latest generation AI processor, the Zhenwu M890, has achieved commercial application through Alibaba Cloud among over 650 external clients across more than 20 industries, including autonomous driving, internet, and finance.

The commercialization process is accelerating. Super-node instances based on the Zhenwu M890 were recently launched on Alibaba Cloud and have begun scaled sales, with Eddie Wu stating that volume will continue to increase in the second half of the year.

During the earnings conference call, Eddie Wu revealed that to date, over 500,000 units of the previous generation of T-Head chips have been produced and shipped. The latest generation of chips was deployed on Alibaba Cloud's AI platform in August, offered as a service in Supernode form. "I believe we are one of the few companies currently capable of deploying such large-scale self-developed domestic chips," he said.

According to Alibaba Cloud, the "Zhen 5 M890" super-node, based on the new generation of T-Head chips, can run inference for large models with over 2 trillion parameters. Kimi K3 and Qwen 3.8 Max are already providing services using these instances.

For Alibaba Cloud, the full-stack layout of self-developed chips covering GPUs, CPUs, and network chips, along with their scaled commercial use and external sales, signifies that T-Head's computing power is moving from internal support to scaled sales, becoming a new growth variable for the cloud business.

Accelerated Data Center Delivery

Efficiency improvements on the computing supply side are also significant. Alibaba Cloud has compressed the delivery cycle for large-scale AI data centers to 100 days and expects the production efficiency of its self-developed modular data centers to more than double in 2026 to meet stronger AI computing demand.

Behind this efficiency gain is upfront capital expenditure. By the end of the first fiscal quarter, Alibaba had invested approximately RMB 190 billion of its three-year, RMB 380 billion investment plan announced in February 2025.

Alibaba Chief Financial Officer Xu Hong stated that AI commercialization is built on infrastructure such as computing centers, noting that "capex must come first to achieve subsequent business growth."

According to their calculations, AI computing assets can recoup costs in about three years. As the gross margin of AI-related products continues to rise, the payback period is expected to shorten further.

Going forward, attention should focus on the ramp-up pace of the Zhenwu M890 super-nodes and whether the profit margin of the AI cloud and computing power segment can continue to rise in subsequent quarters.