Alibaba Group Holding Limited BABA

Key data

The business

Alibaba is three businesses stapled together. The biggest and oldest: Taobao and Tmall, China's dominant online marketplaces, where merchants pay Alibaba for placement and commissions rather than Alibaba holding much inventory. That is the cash machine. Second: Alibaba Cloud, the country's largest cloud provider and now its AI arm, renting computing power and selling its Qwen models. Third, and newest as a profit drain: quick commerce, food delivery and one-hour grocery under Taobao Instant Commerce, Freshippo and Tmall Supermarket, where Alibaba is spending hard to hold share against Meituan and JD. The moat is the marketplace: hundreds of millions of Chinese shoppers and the merchants who must reach them, a two-sided network no rival has dislodged in two decades. The thing a customer touches is the Taobao app they open to buy a phone case; the thing that pays the bills is the merchant buying ads to sell it to them.

The numbers

The story lives in one column of the table: operating income, the profit from actually running the businesses.

Revenue climbed to $148.4B in FY26 (ended March 2026), up about 8% in dollars, low-single digits in the weaker yuan it actually earns. But operating income fell 63%, from $19.4B to $7.3B, and the operating margin went from 14% to 5%. Net income fell far less, down 16% to $15.0B, because investment gains filled the hole, non-repeatable earnings propping a headline number.

The quarterly misses tell the same story from a different angle: three of the last four came in near half the profit forecasters expected, the March quarter at $0.09 per ADS against a $1.02 look. The spending is real and it is landing on the bottom line now.

Here is the whole question in one calculation. A 14% operating margin on $148B of revenue is about $21B of profit; a 5% margin is $7B. The market must decide whether the missing $14B is an investment that returns as cloud and AI revenue, or a toll Alibaba now pays every year to defend commerce against Meituan and JD. Cloud already runs at a 12% operating margin per company disclosure and its growth is accelerating, which argues for the first; the delivery war argues for the second. That is I think the cloud half is worth more than the consolidated profit lets you see, and the print that settles it is China commerce margin holding while cloud growth stays above the high twenties.

Management

Two insiders said opposite things with their own money this summer. Joseph Tsai, the chairman, bought about $20.7M of stock across August 24 and 25, just as the company raised money for AI. Michael Evans, the president, sold $65.8M on June 29, plan status not disclosed. On the company's own account, the tell is the placement itself: 710 million new shares at HK$112.70, about $10B and roughly 4% dilution, with 100% earmarked for AI and 60% of it for data-center computing. Alibaba spent years buying its own stock back, over $11B in FY25; in FY26 buybacks collapsed to $1.1B and it turned around to issue shares instead. That is the clearest signal in the file: management would rather own more AI capacity than fewer shares.

How it fails or surprises you

The delivery war never ends. If China quick-commerce losses stay structural, consolidated operating income sits near FY26's $7.3B instead of recovering toward FY25's $19.4B. Meituan and JD have cash and no reason to retreat. Shows first as China commerce adjusted EBITA down again next quarter, with quick-commerce losses widening rather than the improving unit economics the CFO claimed.

Cloud and AI re-rate the whole company (right tail). Cloud already earns a 12% operating margin per company disclosure and its growth is accelerating; the $10B raise buys the data centers to feed it. If cloud revenue growth pushes past 30% with margins rising, that segment alone justifies today's price, and it sits unpriced inside a 5% consolidated margin. Shows as two straight quarters of cloud growth above the high twenties.

The raise is not the last one. Free cash flow already turned negative, with capital spending near twice operating cash flow. If AI outlays keep outrunning cloud money coming in, the $10B placement becomes the first of several, and each one dilutes. Shows as next-quarter capital spending against cloud revenue, and any second placement within a year.

Closing thoughts

The market sees a 5% operating margin and a stock down 41% from its 52-week high, and treats the whole thing as a broken compounder. What it is not separately pricing is a cloud business at a 12% margin and accelerating, hidden inside the consolidated wreckage. Over the next two quarters two numbers settle it: China commerce margin, and cloud growth. If commerce margin steadies while cloud stays above the high twenties, the missing profit was an investment and earnings recover; if commerce margin keeps sliding, it was a toll and 25 times earnings is dear on real profit. The downside tail is fatter in the near term, because the delivery war is live and the dilution is certain; the upside is worth more if it lands, because nobody is paying for cloud today.

The bet is still that Alibaba can build its AI machine and fight its delivery war without breaking the shopping-and-cloud engine that pays for both. The one pair of numbers that tells you first is China commerce margin against cloud growth, quarter by quarter. The bet is wrong the day operating income is still near $7B a year from now with more shares outstanding than today.