news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Produced by Huxiu Business Consumption Group

Author: Miao Zhengqing

Title image｜Visual China

Tiger Sniff Note: This article is the 002nd article of "What's Happening in Silicon Valley", which focuses on foreign leading AI-related companies such as Open AI, Microsoft, and NVIDIA. This article starts with Microsoft's latest financial report: Is the impetuous Wall Street willing to wait for Microsoft's "15-year plan"?

Microsoft's ambitious investment in AI and the outside world's suspicion constitute a tense picture. The subtle "pull" emotion permeated Microsoft's earnings conference call in the early morning of July 31.

In the early morning meeting, facing questions from a group of analysts, Microsoft Chairman and CEO Satya Nadella and Microsoft Executive Vice President and Chief Financial Officer Amy Hood spoke in succession, trying to reassure investors and convince them that Microsoft's huge investment in AI is reasonable and expected to pay off.

Earlier in the day, Microsoft released its quarterly financial report ending June 30. Microsoft's operating income and operating profit both increased by more than 15% year-on-year to $64.7 billion and $27.63 billion, respectively. However, due to factors such as increased investment, Microsoft's net profit growth rate is slowing down, from a year-on-year increase of 20% in the previous quarter to a year-on-year increase of 10% to $22 billion.

In this quarter's financial report, Microsoft's large investment in AI is impressive:Microsoft's capital expenditures in the quarter reached $19 billion, the highest in a single quarter since fiscal 2024, while the previous quarter's spending was about $14 billion. In other words, in the past six months, Microsoft's capital expenditure was about $33 billion, which exceeded the total GDP of Iceland in 2023.

Microsoft's heavy investment is mainly in AI, especially the infrastructure of AI. Amy Hood said in a conference call that capital expenditures are almost all used for AI,cloud computing, and half of it is used for infrastructure investment (such as some data centers and sites for building centers), which are built mainly to meet AI-related needs.

What makes investors doubtful is the certainty and return cycle of huge investment returns.

According to Microsoft's prediction, this will be a future return of more than 10 years. "These investments are crucial for AI, but it may take 15 years or even longer to get a return," said Amy Hood.

Shortly after this statement, Microsoft's US stock began to fall, falling more than 2.58% from the day's high, and then narrowed to 1.19%. By the close of the day, Microsoft's US stock fell by about 0.89% to $422.92 per share.

Another factor that makes investors uneasy is that among the various revenues announced by Microsoft this time, Microsoft did not disclose the specific revenue of the AI ​​tool product "Copilot for Microsoft 365", which has attracted much attention from the outside world. In sharp contrast, Microsoft described the situation of its Azure cloud service business "meticulously and carefully" in its financial report and conference call, revealing not only the changes in revenue, but also the use of Azure's non-AI data centers in Europe.

Therefore, even the most dull observers and analysts will realize that Microsoft’s investment in AI is currently heavy and will continue to increase; it is difficult for Microsoft to make a living by relying on AI.

Later in the day, DADAVIDSON and Morgan Stanley both cut their price targets on Microsoft.

Two unfavorable factors have become a shadow for Microsoft

The cloud computing performance fell short of expectations and the heavy investment in AI, which is expected to last up to 15 years, has made some investors skeptical about Microsoft's prospects.

Let’s first look at the cloud computing business.

According to data released by Microsoft, its Azure cloud service business revenue increased by 29% year-on-year, lower than the previously expected growth rate of 30.6%. It is worth noting that over the past year, Microsoft's quarterly growth rate of Azure cloud service business revenue has exceeded 30%.

What further affects the confidence of the outside world is that based on the quarterly results, Microsoft has also adjusted the revenue growth rate of its Azure cloud service business in the next quarter. The latest information released by Microsoft shows that the expected growth rate is 28%~29%, which is slightly different from the previous expectation of 29.7%. In short, the growth of Microsoft's Azure cloud service business is slightly weaker than the outside world expected.

The outside world's expectations for Microsoft's Azure cloud service business essentially come from "AI". Among Microsoft's major business lines, Azure cloud service business is the business that is most closely integrated with AI. It can even be said that Azure cloud service business plays the role of a "barometer" of whether Microsoft's AI efforts are going well.

Microsoft's Azure cloud service business not only provides the "computing power base" for AI products, but also integrates Microsoft's heavily investedOpenAISome of its product features.

In Microsoft's AI layout, its external AI investment, internal AI product research and development, and Azure sector are working together: the Azure cloud business not only plays the role of a computing power base, but also provides a "computing power payment" function. On this basis, Microsoft expands the AI ​​radiation radius by investing in companies such as OpenAI; internally, Microsoft completes the AI ​​transformation of the entire product line by embedding OpenAI's technology and some self-developed technologies into its own products.

This is also why Microsoft's stock price fluctuates when the growth rate of Azure cloud service business is lower than outside expectations.

It is worth noting that the signal released by Microsoft this time is that the revenue growth rate is lower than expected. However, the growth of Azure cloud service business on the user side has increased significantly. According to Satya Nadella in the conference call, there are currently more than 60,000 Azure AI customers, with a year-on-year growth rate of 60%.

But the challenge facing Microsoft is whether these new AI users can contribute the "AI consumption power" that Microsoft executives expect - to use or even pay for the various AI tool SKUs provided by Microsoft.

"At the end of the day, AI is just software," Nadella told analysts at a meeting that evening.

But Microsoft still has huge room for improvement in this key link: as of the latest quarter, AI-related services contributed 8% to Azure's overall revenue, compared with about 7% in the previous quarter.

Another point that has deepened the outside world's doubts about Microsoft is that Microsoft's thinking on AI investment is tending to be "very physical." In the long run, this seems to be the only way to go in the AI ​​track, but for Wall Street, which has been spoiled by the "FAANG" myth, this plot of "taking a long time to get a big return" is too procrastinated.

According to Amy Hood, most of Microsoft's capital expenditures are used for data center construction and sites, and the rest are invested in servers (including GPUs and CPUs). In short, Microsoft is trying to further strengthen its advantages in "infrastructure and foundation" in this AI boom.

However, the monetization cycle of these heavy investments is not only long, but also "currently there is limited room for imagination." Currently, among the AI ​​products that Microsoft can really monetize, the most famous one is "Copilot for Microsoft 365" - an AI product that charges users a monthly fee.

"Copilot for Microsoft 365" may foreshadow the outline of the route of Microsoft's subsequent AI products: OpenAI, which is invested by Microsoft, will provide the main technology, Azure will provide the computing power base, and the functions or services of the new products will be connected with other products.

But one test comes from the ability to monetize: as of now, the monthly fee for "Copilot for Microsoft 365" is US$30/month, and Microsoft's capital expenditure on AI and cloud computing has exceeded US$33 billion in the past six months.

In addition to the difficulty of monetizing AI, Microsoft also has constant small troubles.

Just a dozen days before Microsoft released its earnings report, Microsoft Azure had just experienced a "major outage". On July 18, some Microsoft Azure users encountered a malfunction, and the "blue screen" became a global hot search. At the time Microsoft released its earnings report, the malfunction in some areas had not yet been completely eliminated.

In the AI ​​field that Microsoft is obsessed with, Microsoft has been sued several times this year. In April this year, Microsoft was sued by some overseas media and accused of "abusing media results to train generative AI." In July, Microsoft faced an antitrust investigation in the UK due to its cooperation with Inflection AI.

This is Microsoft's current situation: it is trying to complete an "elephant dance" in the AI ​​track, but this will be a performance that will not begin until more than ten years later. Before that, Microsoft can only wait patiently and try to reduce the number of "fleas" on its body as much as possible.