Alibaba Cloud Begins to Profit from AI

11/22 2026 562

Text / Liang Tian

Source / Node AI

Over the past two years, a rather interesting phenomenon has emerged in the tech industry.

Nearly every major company is talking about AI, and nearly every tech giant is ramping up investment in computing power. But ask one more question—how much money has AI actually made for you?—and many companies struggle to answer.

Some rely on model financing, some rely on applications to attract users, and some are busy buying GPUs and building data centers. The AI industry looks booming, yet far fewer companies than one might imagine can truly turn technological enthusiasm into revenue—and then turn revenue into profit.

On November 19, Alibaba released a financial report worth scrutinizing.

The group's quarterly revenue reached 271.519 billion yuan, up 10% year on year. Compared with the group's overall performance, Alibaba Cloud's numbers are more noteworthy.

In this quarter, Alibaba Cloud's revenue from AI cloud and computing power services reached 60.973 billion yuan, up 53% year on year; revenue from external customers increased by 64% year on year; adjusted EBITA reached 8.838 billion yuan, up 121% year on year, with the profit margin rising to 14.5%.

There is another figure: AI-related product revenue has achieved triple-digit year-on-year growth for 13 consecutive quarters. Taken together, these numbers send a signal worth attention: Alibaba Cloud is not only growing—that growth is improving its profitability.

Node AI believes that the most discussion-worthy aspect of this financial report is not how much Alibaba Cloud has grown again, but that a shift is emerging in China's AI industry: commercial competition in AI is beginning to move from whose technology is more advanced to who can truly turn technology into business.

Why Has Alibaba Cloud Suddenly Sped Up?

To understand this round of growth at Alibaba Cloud, we must first understand one change: the reasons enterprises buy cloud services today are already somewhat different from a few years ago.

In the past, enterprises moved to the cloud mainly to save IT costs.

They previously had to purchase servers themselves, build computer rooms, and maintain systems. Now, by handing this work to cloud vendors, they can both reduce initial investment and improve operational efficiency.

This was an infrastructure business.

But after the emergence of AI, enterprise demand for cloud began to change.

Training large models requires computing power, deploying models requires computing power, and enterprises using agents to handle business also require computing power.

More importantly, a model is not finished once training is complete. As long as applications continue to run, inference demand will arise.

In the past, enterprises bought cloud more to make their existing business run more cheaply; now enterprises buy AI cloud in the hope of creating new business value through it.

These two types of demand define different market spaces.

There is a data point in Alibaba's financial report that is easy to overlook: recurring revenue from external customers (excluding revenue from computing power equipment solutions) increased by 67% year on year.

This means that this quarter's growth does not rely solely on one-time equipment solution sales—a considerable portion comes from sustained cloud service demand.

Of course, growth in recurring revenue does not necessarily mean customers will definitely stay long term, but at least it shows that AI demand is entering Alibaba Cloud's revenue structure in a more sustained way.

Alibaba disclosed that customers in industries such as internet, automotive, embodied intelligence, financial services, and biotechnology are all expanding their use of AI cloud services. This is more noteworthy than simply serving a few large model companies. Because the real commercial space for AI lies not only in model companies, but in those enterprises willing to continuously pay for AI capabilities.

More Important Than Revenue Growth Is That Profit Has Begun to Catch Up

Cloud computing has a rather cruel characteristic: large scale does not necessarily mean profitability.

Data centers must be built, servers must be purchased, chips must be updated, and electricity and operations and maintenance all cost money.

Especially in the AI era, the threshold for computing power investment is getting higher and higher.

Therefore, to evaluate an AI cloud vendor, one cannot look only at revenue growth rate, but also at whether growth can cover increasingly heavy investment.

What is noteworthy about Alibaba Cloud this time is precisely here.

Revenue increased by 53% year on year, while adjusted EBITA increased by 121%.

That is to say, in this quarter, Alibaba Cloud's profit growth was significantly faster than its revenue growth.

This at least shows that the expansion of Alibaba Cloud's business scale has begun to reflect better efficiency at the operating profit level.

This may include multiple factors, including an increased share of high-value AI services, improved resource utilization efficiency, and economies of scale. But the financial report does not fully break down the profit contribution of each factor, so the improvement in profit cannot simply be attributed entirely to AI.

There is another account that must be noted.

In this quarter, Alibaba Group's free cash flow showed a net outflow of 34.472 billion yuan, compared with a net outflow of 21.840 billion yuan in the same period last year. The company explained that the main reason was increased spending on cloud infrastructure.

This means that although Alibaba Cloud has performed well at the operating profit level, the infrastructure supporting future growth still requires a large amount of real money. This is also something worth continuing to observe in the coming quarters.

What Alibaba Cloud Really Wants to Do May Be More Than Selling Computing Power

If Alibaba Cloud is seen as a company selling server resources, it is hard to understand why it simultaneously invests in chips, large models, and AI applications.

But from another angle, this logic becomes clear.

Alibaba is trying to connect a complete AI industry chain: at the bottom are Pingtouge chips, in the middle are cloud computing infrastructure and the Qwen large model, and at the top are applications such as Qwen Office.

Chips determine computing power supply, models determine intelligence capabilities, applications determine user demand, and the cloud platform is responsible for organizing these capabilities and delivering them to customers.

For example.

An e-commerce company hopes to use AI to manage its store. In the past, it might have needed to separately purchase model services, data processing tools, and business software, and then complete system integration itself.

Now Alibaba's Qwen Office e-commerce edition has begun to connect store operations, business analysis, marketing, and customer service, and supports data access from Taobao and Tmall.

This means that AI is no longer just answering questions raised by merchants, but attempting to directly participate in merchants' business processes.

For Alibaba, this matter has another layer of significance.

When enterprises use AI applications, model calls, data processing, and cloud computing demand may be generated simultaneously behind the scenes. The deeper the applications go, the greater the opportunity to drive sustained consumption of underlying services.

This is a business model worth watching: no longer simply waiting for customers to buy computing power, but creating computing power demand by solving business problems.

Node AI believes that in the future, competition among AI cloud vendors will not only be competition over computing power prices, but also competition over understanding of enterprise business. Whoever can truly bring AI into enterprise business processes will have the opportunity to turn one-time technology procurement into long-term commercial relationships.

Why Is Alibaba Cloud Getting Heavier and Heavier?

Alibaba Cloud's report card is good, but one reality cannot be ignored: AI cloud is a business that is getting heavier and heavier.

In this quarter, Alibaba Group's capital expenditure reached 50.477 billion yuan, up 60% year on year.

The company disclosed that the increase in spending is related to factors such as the expansion of computing power scale, rising prices of chip components, and growth in CPU demand brought by agents.

More noteworthy is that the competition rules of the AI industry are still changing rapidly.

On the one hand, model capabilities continue to improve, and inference costs may continue to decline. On the other hand, the spread of applications may bring more computing power demand.

Can cost declines translate into sufficiently large growth in usage? Can scale expansion continue to improve profit margins? These questions currently have no final answers.

However, for customers, performance, price, stability, security, and migration costs are always real procurement criteria.

Therefore, Alibaba Cloud's full-stack layout may become a competitive advantage, but it also means it must bear relatively high research and development and infrastructure investment over the long term.

Fortunately, Alibaba has ample capital reserves. As of the end of September, the group had approximately 487.9 billion yuan in cash and other current investments, providing some room for subsequent investment. That said, ample funds do not mean unlimited investment. As AI infrastructure becomes more and more expensive, return on capital will ultimately become an important indicator for testing strategy.

Final Words

Looking back at Alibaba Cloud's financial report, I think what is truly valuable about it is that it lets the market see a more concrete side of AI commercialization. AI does not only tell stories about the future; it has also begun to contribute current-period revenue and operating profit.

Quarterly AI-related product revenue of 17.918 billion yuan shows that demand is taking shape at scale; adjusted EBITA growth of 121% shows that profitability is improving; and continuously increasing capital expenditure reminds us that this competition is far from entering an easy harvest stage.

Therefore, to judge the value of an AI company, one cannot look only at how smart the model is or how massive the computing power is, but also at three questions: whether customers continue to pay, whether it can improve customers' operating efficiency, and whether the money earned can cover long-term investment.

None of these three accounts can be missing.

Alibaba Cloud has at least delivered stage results meeting expectations on the first two accounts.

Solemnly declare: the copyright of this article belongs to the original author. The reprinted article is only for the purpose of spreading more information. If the author's information is marked incorrectly, please contact us immediately to modify or delete it. Thank you.