Alibaba's AI story is shifting from promise to profit—but it's costing a fortune to get there.

CEO Eddie Wu Yongming told investors this week that the company may blow past its planned AI capital spending as it goes all-in on a "full-stack" artificial intelligence strategy. That means pouring money into chips, cloud infrastructure, models, and applications all at once. And the market is paying close attention.

Alibaba said annualized recurring revenue from AI models and applications could reach 30 billion yuan ($4.42 billion) by the end of the year. That's a big number, but it comes with a big price tag. The company's adjusted EBITA fell 84% in the fiscal fourth quarter as it ramped spending on AI, cloud expansion, and e-commerce upgrades. Revenue rose just 3% year over year to $35.28 billion, and adjusted earnings per ADS missed expectations.

Still, the cloud business is humming. Alibaba's Cloud Intelligence Group delivered 38% revenue growth, marking the eleventh straight quarter of triple-digit growth in AI-related product revenue. Wu said AI-related products now account for 30% of the cloud unit's external customer revenue and could contribute more than half of cloud-computing revenue within the next year.

"Alibaba's AI business has moved beyond the early investment stage into scaled commercialization," Wu said, adding that model and application services are becoming key growth drivers for cloud profitability.

The company also expanded its Qwen AI ecosystem, launched new merchant automation tools, and continued scaling AI-powered shopping features across Taobao and Tmall. Core commerce revenue slipped 1%, but management said growth would have been stronger without merchant incentives and subsidies.

Analysts See AI as the Next Big Engine

Citi analyst Alicia Yap projects Alibaba's AI revenue could grow at a 90% compound annual growth rate from fiscal 2026 to 2031 and eventually contribute 70% of total cloud revenue. That's a stunning trajectory if it holds.

Barclays analyst Jiong Shao told CNBC that both Alibaba and Tencent posted stronger underlying results than headline numbers suggested. He noted Alibaba's China e-commerce business delivered adjusted growth of about 8% year over year, while Tencent's gaming growth was temporarily affected by holiday timing shifts.

Shao emphasized that AI spending has become the key battleground between Chinese and U.S. tech giants. He highlighted Alibaba's plan to spend roughly 380 billion yuan on capital expenditures over three years, with the possibility of even higher spending ahead. He also said Alibaba's cloud business has sustained 30%-40% annual growth for nearly three years and estimated the company's AI-related recurring revenue at around 10 billion yuan annually.

The Chip Problem

But there's a catch. China's biggest AI challenge remains limited access to advanced computing power due to restrictions on high-end chips from NVIDIA Corporation (NVDA). Shao warned that this constraint could slow progress.

He said Alibaba's large AI infrastructure investments are pressuring margins in the short term but position the company as one of China's few true hyperscale cloud and AI competitors globally. In other words, the spending is painful now, but it's building a moat.

What's Next

Alibaba shares were down 3.05% at $141.36 during premarket trading on Thursday, according to market data. The market is clearly weighing the near-term profit pain against the long-term AI opportunity.

For investors, the question is whether Alibaba can sustain its AI momentum while managing the cost of the buildout. Wu seems confident the answer is yes—and he's willing to spend whatever it takes to prove it.