Alibaba Core Profit Plunges 84% as AI and Cloud Capex Bites

Alibaba (NYSE: BABA) had terrible results on Wednesday. The company's main measure of profitability, adjusted EBITA, fell 84% year-over-year to 5.1 billion yuan ($750.9 million) in the March quarter. Revenue grew by only 3% year over year, which was less than what analysts had predicted.

U.S.-listed stocks went up at first in premarket trading, but then they went down as much as 4% as buyers realized how bad the margin compression was.

The numbers in the headlines are shocking. But the bigger picture is important: Alibaba's leaders are making it clear that this is the painful part of a multi-year plan to build out AI infrastructure, and the cloud division's results show that the investment is at least partly paying off.

What Drove the 84% Profit Collapse

Alibaba's margins are being squeezed by two spending programs at the same time. The unusually large difference between top-line and bottom-line success is due to the combined weight of these programs.

The main culprit is the high cost of AI and cloud technology. Alibaba has been putting a lot of money into chips, data centres, and its Qwen family of AI models. These are investments that bring in money right away, but the money keeps coming in over time.

During the earnings call, Alibaba's CEO Eddie Wu said that the company may need to spend more on computer power over the next five years than the $55.9 billion it had planned to spend over the next three years. This suggests that the spending cycle has not yet reached its peak.

The company also said that self-developed AI chips were a strategic goal and that they give them more control over computing supply chains "amid constrained global availability." This was a direct reference to the global shortage of memory and advanced chips caused by AI demand.

The second big drag is tax breaks for quick trade. It costs a lot to advertise in the instant retail area, which offers fast delivery services, because there are a lot of Chinese competitors in this very competitive market.

Quick-commerce sales went up 57% year-over-year, showing that the segment is growing quickly. However, the China e-commerce division's profits are being hurt by this growth. The adjusted EBITA for online shopping in China fell 40% from the previous year. The group's biggest source of income, customer management, only went up 1%, but China e-commerce as a whole went up 6%.

The Bright Spot: Cloud Is Accelerating

Alibaba's cloud business sent the most positive message of the quarter. Cloud-adjusted EBITA went up 57%, and revenue rose 38% year-over-year to $6.1 billion, which was faster than the previous quarter's rate. CFO Toby Xu stated that sales of AI-related products grew by more than 10% for the 11th quarter in a row.

This shows that demand is real and ongoing, not just happening sometimes. In the quarter, AI-related sales brought in a total of $1.32 billion.

The advice for the future is clear and important. Wu said that Alibaba thinks its recurring annualized income from AI model and application services will be more than $1.47 billion in the June quarter and $4.4 billion by the end of the year. This is more than three times what it is now in just one calendar year. If that trend continues, the cloud division's contribution to the group's total profitability will start to change the story of the margin in a big way in 2026.

It was easy for Wu to explain the expected return on investments: "Returns on the company's investments over the next three to five years are extremely clear." While the spending cycle is going on, Alibaba wants investors to believe in this message: we are sure of where we want to go and honest about when we can get there.

The Global AI Capex Pattern Alibaba Fits

Alibaba's margin compression is not unique; it is the buy-side version of a pattern that can be seen in all of the world's big AI infrastructure investors. There is a wave of capital expenditures that is good for Nvidia, Broadcom, SK Hynix, and TSMC, but it is bad for Alibaba's income statement.

Investors in the four biggest hyperscalers in the U.S. are facing the same earnings-versus-investment stress because they all have $700 billion planned for capital expenditures in 2026.

The thing that makes Alibaba different is that it is spending a lot on AI infrastructure at the same time as a domestic quick-commerce support war, and its sales are only growing at 3% per year. The 84% drop in EBITA was caused by three things: AI capital expenditure, advertising spending, and slower top-line growth. This is different from the normal 20–30% margin pressure that happens during pure-play AI investment cycles.

The planned short-term cost of a conscious long-term bet is an 84% drop in core profit. The 38% increase in revenue and 57% increase in EBITA for the cloud division back the idea that AI is in demand and that Alibaba's investments in infrastructure are paying off. The $4.4 billion revenue goal for AI by the end of the year gives management a clear goal to work toward and can be used to judge their trustworthiness.

Time is the risk. Wu's "three to five years" time frame for full investment returns means that present shareholders will have to deal with multiple quarters of margin pressure before they get their money back.

If AI sales slow down or the competitive situation in quick-commerce doesn't return to normal, the road from today's 84% drop in profits to tomorrow's AI making money could be longer and cost more than what is currently expected. The next real test of whether the ramp is going as planned will be the AI revenue number for the June quarter, which is expected to be over $1.47 billion a year.