Alibaba’s AI Cloud Is Growing Fast. Its AI Bill Is Growing Faster.

Alibaba’s AI Cloud and Compute Services revenue rose 45% to 48.44 billion yuan in the June quarter, but the company’s capital expenditure climbed 75% to 67.68 billion yuan and adjusted earnings missed analyst expectations. The results put a hard number on the tradeoff in Alibaba’s AI strategy: strong demand is funding growth, while the infrastructure required to serve it is pressuring near-term returns.

Total quarterly revenue rose 9% year over year to 268.95 billion yuan, broadly in line with the 268.88 billion yuan estimate cited by LSEG. Adjusted earnings were 8.52 yuan per American Depositary Share, below LSEG’s 10.53 yuan estimate, while net income fell 75% from a year earlier.

Alibaba attributed much of the quarter’s profitability pressure to increased technology investment. Adjusted EBITA, a measure of operating profit before interest, taxes and amortization, declined 30% to 27.3 billion yuan; its margin fell to 10% from 16% a year earlier. Free cash flow moved to negative 44.7 billion yuan, compared with positive 18.8 billion yuan a year earlier.

The 45% growth rate reflects revenue from Alibaba’s cloud and AI computing business, while AI-related products contributed 12.4 billion yuan during the quarter. Alibaba said AI-related product revenue had posted triple-digit year-over-year growth for 12 consecutive quarters. CEO Eddie Wu said the company benefited from improving commercialization of its full-stack AI capabilities.

Alibaba has changed how it shows the AI operation

• Alibaba Cloud and its T-Head chip-design business are reported together as AI Cloud and Compute Services.
• Qwen Consumer Business Group and Qwen Work are reported under AI Labs and Applications.
• Alibaba revised its quarterly reporting to provide more detail on AI-related activities.

Alibaba has used roughly half of the 380 billion yuan it plans to invest in AI infrastructure through 2029, according to the cited account. Wu said the AI spending could break even within three years based on current gross margins, an outlook that depends in part on expanding use of Alibaba’s in-house T-Head chips rather than commercially purchased processors.

Alibaba said its T-Head chips are already running at scale on linked server racks known as supernodes. That provides a tangible part of management’s margin argument, but the company has not yet reached the forecast break-even point for its AI spending. For investors, the next test is whether cloud and AI-product growth can continue while the capital bill and margin pressure ease.