Alibaba's $10 Billion AI Gamble: Profit Crashes 75% as Cloud Surges 45%

On August 20, 2026, Alibaba Group delivered the clearest picture yet of what the AI infrastructure race actually costs — and what it might eventually pay. The Chinese tech giant reported that net income in its June quarter collapsed 75% year-over-year to roughly RMB 10.54 billion (about $1.55 billion), a plunge driven almost entirely by its aggressive spending on AI data centers, chips, and computing capacity. At the same time, Alibaba Cloud’s external revenue growth accelerated to 45% year-over-year, its fastest pace in years, with AI-related products now generating a run-rate of RMB 49.5 billion annually.

The market’s reaction was immediate and unforgiving: Alibaba’s US-listed shares fell around 5% in volatile premarket trading. Investors, it seems, can read an income statement. But the more interesting story is buried in the balance sheet strategy — whether this quarter marks the peak of the pain or merely the end of the beginning.

The numbers behind the plunge

The June quarter — Alibaba’s fiscal first quarter of 2027, though most outlets are filing it under the 2026 calendar year — tells a story of two companies operating inside one. On the profitability side, the picture was bleak: net income down 75% (Quartz and the Wall Street Journal round it to 76%) to RMB 10.54 billion, or roughly $1.55–1.6 billion. On the growth side, the picture was the opposite: total cloud revenue rose 34% year-over-year, external cloud revenue growth accelerated to 45%, and overall quarterly revenue beat analyst estimates.

The bridge between those two facts is capital expenditure. Capex for the quarter came in at RMB 67.7 billion — approximately $9.5–10 billion — driven almost entirely by AI infrastructure investment. That is a 75% increase from the year-ago quarter, a pace of spending that overwhelms even fast-growing operating profits. For context, Alibaba’s full fiscal 2026 capex ran roughly $18.3 billion across four quarters, averaging about $4.6 billion per quarter. The June quarter alone nearly doubled that average.

The company is also compressing its spending timeline. Management indicated the pace of deployment means its planned AI investment horizon is shortening — from three years toward 2.5 years or less. CEO Eddie Wu, who has staked his tenure on the AI transition, said the company is in a “superior position” for the AI boom, and reiterated that AI-related product revenue is expected to eventually account for 50% of Alibaba Cloud’s external revenue, up from roughly 30% today.

The 380 billion yuan commitment

This quarter is not an anomaly — it is the execution of a plan announced more than a year ago. In early 2025, Alibaba committed to investing RMB 380 billion (over $50 billion) in AI and cloud infrastructure over three years, one of the largest single corporate technology commitments in China’s history. By May 2026, Wu had already signaled the company would “likely overshoot” that figure. The June quarter’s RMB 67.7 billion capex puts the company on an annualized run-rate of roughly RMB 270 billion — ahead of the pace needed to hit, and likely exceed, the original target.

Where is the money going? AI data centers, GPU and custom accelerator fleets, and the networking and power infrastructure that modern training and inference clusters demand. The Bloomberg segment on “circular AI financing” points to a growing concern among analysts: much of this spending is interconnected — cloud providers funding compute that is consumed by AI labs whose valuations depend on AI demand, which in turn drives more infrastructure buildout. Whether that loop is a virtuous cycle or a bubble in formation is the defining question of the current phase of the AI boom, and Alibaba’s income statement is now the cleanest real-world data point in that debate.

Cloud growth: the payoff side of the ledger

If capex is the cost, cloud revenue is the return, and here Alibaba has genuine momentum. External cloud revenue growing 45% year-over-year is a dramatic acceleration — the same metric grew 40% in the March quarter, and 38% before that. AI-related products now account for roughly 30% of cloud external revenue and reached a RMB 49.5 billion annualized figure in the June quarter, per MarketBeat’s earnings summary.

That growth is being driven by Qwen, Alibaba’s open-source model family, which has become one of the most downloaded and fine-tuned model lines in the world. The strategic logic is straightforward: open-weight models drive developer adoption, developer adoption drives cloud consumption, and cloud consumption monetizes the infrastructure Alibaba is building at a loss today. It is the AWS playbook updated for the generative AI era — with the added twist that Qwen’s openness makes it a genuine export product, spreading Alibaba’s AI stack across Southeast Asia, the Middle East, and beyond.

The question is whether 45% cloud growth can outrun 75% capex growth. Right now, it cannot — that is precisely why profit fell 75%. The bet is that infrastructure spending is front-loaded while revenue compounds. If AI demand persists, the gap narrows from both directions. If it does not, Alibaba will own a lot of very expensive servers.

The competitive and geopolitical context

Alibaba’s earnings land in the middle of China’s broader AI capex boom, which analysts now estimate at roughly $295 billion in planned data center and compute investment. Baidu reported its own mixed quarter days earlier — revenue down 2%, but AI-related income up nearly 50% to RMB 5.6 billion — showing the same pattern of AI revenue growing fast off a small base while legacy businesses stagnate.

Geopolitically, the spending surge is happening against the backdrop of intensifying US–China tech competition. China has urged respect for digital sovereignty in the AI race even as Washington pushes its “Pax Silica” coalition partners to choose sides. Domestic chips and sovereign compute are national priorities, and Alibaba — with its custom silicon work and its role as China’s largest cloud provider — sits at the center of that push. Its willingness to absorb a 75% profit decline is partly commercial calculus and partly a reflection of how strategically vital AI infrastructure has become in Beijing’s eyes.

What to watch

Three signals will determine whether this was a smart bet. First, cloud growth sustainability: 45% is spectacular, but the market will want to see it hold or accelerate in the September and December quarters. Second, AI revenue mix: the march from 30% of cloud revenue toward Eddie Wu’s 50% target is the single clearest indicator that demand is real. Third, capex trajectory: management’s hint that the investment horizon is compressing suggests spending will stay elevated through at least 2027 — meaning more painful quarters ahead before the picture improves.

For the broader industry, Alibaba’s quarter is a stress test being run in public. Every hyperscaler — Microsoft, Google, Amazon, and Meta among them — is making a version of the same trade: sacrifice current margins for AI capacity. Alibaba is simply doing it with more violence and less ambiguity than anyone else. The 75% profit drop is the price of admission; the 45% cloud growth is the prize. Which number ends up defining the era is the question every AI investor is now asking.