Alibaba's AI Gamble: Profit Crashes 75% as Cloud Revenue Rockets 45%

Alibaba reported its June quarter 2026 results on August 20, and the numbers read like a controlled demolition of its own income statement. Group revenue rose a modest 9% year-over-year to RMB 268.95 billion (US$39.64 billion). But GAAP net income cratered 75%, group adjusted EBITDA fell roughly 30% to RMB 27.3 billion, and earnings per ADS of US$1.26 missed analyst estimates by about $0.34. The culprit was not weakness in the business — it was the single most aggressive capital spending program in the company’s history, aimed squarely at artificial intelligence.

The headline numbers

The quarter’s most striking figure sits in the Cloud Intelligence Group. External cloud revenue grew 45% year-over-year to RMB 48.44 billion (roughly US$7.1 billion), a sharp acceleration from the 38% growth posted in the March quarter and far above the 26% recorded a year ago. AI-related product revenue continued its extraordinary streak — twelve consecutive quarters of triple-digit year-over-year growth, reaching RMB 12.4 billion for the quarter. On the earnings call, CEO Eddie Wu said growth would “further accelerate in the coming quarters,” a claim few companies on earth can make at this scale.

That acceleration came at a price. Capital expenditure surged 75% year-over-year to nearly US$10 billion for the quarter alone, as Alibaba expanded data center capacity to meet computing demand it says it still cannot fully satisfy. Free cash flow has already turned negative — the company reported a RMB 46.6 billion outflow for fiscal 2026 — and management has signaled it will blow past its February 2025 pledge of RMB 380 billion (about US$56 billion) in AI and cloud infrastructure investment over three years. One analysis puts the net loss across Alibaba’s AI operations at RMB 8.2 billion for this quarter alone.

Why the market shrugged

Here is the paradox: despite the profit collapse and the earnings miss, Alibaba’s shares jumped as much as 8% in US trading — well above the roughly 7% move options markets had priced in. Investors are no longer valuing Alibaba as an e-commerce company with a tech side project. They are valuing it as China’s answer to the American hyperscalers, and the 45% cloud growth number is the evidence they were waiting for.

The market’s logic is straightforward. Cloud capacity in China is reportedly sold out or nearly so, with demand for AI training and inference exceeding supply. Every yuan of capex Alibaba deploys today converts into contracted cloud revenue tomorrow — the 45% growth rate is mechanically tied to the compute being switched on. Management has also set an explicit target of surpassing US$100 billion in combined external cloud and AI revenue over five years, and model-plus-application services ARR is expected to pass RMB 10 billion annualized in this quarter and RMB 30 billion by year-end. Those are the metrics the market is now watching, not quarterly EPS.

There is also a strategic dimension. Alibaba’s Qwen model family — most recently expanded with its biggest model yet, claiming performance on par with Anthropic’s frontier offerings — has become the default open-weight ecosystem in China and increasingly elsewhere. Open-source Qwen drives developers onto Alibaba Cloud, which drives revenue, which funds more capex. The flywheel only works if the infrastructure exists, which explains management’s willingness to let profit fall 75% rather than throttle spending.

The circular financing question

Not everyone is celebrating. The same day as the earnings, Bloomberg flagged growing concern about “circular AI financing” — the web of deals in which AI vendors, cloud providers, and investors effectively fund each other’s growth, inflating apparent demand. Combined 2026 capital spending by Alphabet, Microsoft, Meta, and Amazon is tracking toward roughly US$700 billion, up more than 60% from last year’s record, and total AI spending is estimated at US$2.5 trillion for the year. Critics, including prominent investors, warn that if end-user demand disappoints, the circularity could unwind violently.

Alibaba sits inside this debate but is not purely a creature of it. Its cloud revenue comes overwhelmingly from enterprises renting real computing power for real workloads, and its core commerce business — though customer management revenue slipped 7% amid China’s consumption funk and instant-commerce price wars — still generates the cash that funds the buildout. The company’s balance sheet, with a large net cash position, gives it a cushion that pure-play AI infrastructure startups lack.

What to watch

Three things will determine whether this bet pays off. First, whether cloud growth stays above 40% as new capacity comes online through the rest of the calendar year — deceleration would be read as demand softening. Second, whether AI product revenue keeps its triple-digit streak as inference increasingly replaces training as the dominant workload. Third, how quickly the RMB 30 billion model-and-application ARR target is reached, since that is the clearest signal of real enterprise adoption rather than mere capacity rental.

The June quarter will be remembered as the moment Alibaba stopped hedging. A 75% profit drop would once have triggered a shareholder revolt; in August 2026, it triggered a rally. Whether that reflects wisdom or mania depends entirely on whether China’s AI demand curve keeps bending upward — and on that question, Alibaba has now bet roughly a hundred billion dollars.