Microsoft's cloud sales are disappointing because it is difficult to meet AI demand
Free unlocking editor's abstract
FT's editor Roula Khalaf chose her favorite story in the weekly news and communication.
The growth rate of Microsoft's vast cloud department is slower than the forecast of Wall Street, because the technology group strives to maintain the pace of customer needs for artificial intelligence related services.
The Software Group headquartered in Seattle defeated analysts' expectations for this quarterly revenue and net income in December, but its cloud department (its maximum income driving force, including its Azure Cloud Computing Platform).
Although AZure's AI service revenue increased by 157 % a year, Amy Hood, the chief financial officer, said that the “continuous impact” of Microsoft's expectation ability restrictions continued in 2025 because the company is committed to “cope Execution of challenges. ” She said this will be improved at the end of the fiscal year.
The overall cloud department of Microsoft increased by 21 % a year ago to $ 40.9 billion, which lost US $ 41.1 billion in the Bloomberg survey.
The disappointing results of the cloud department made the stock fell more than 4 % during the transactions after get off work in New York, which made the group's market value of about $ 150 billion.
The collective revenue of this quarter increased by 12 % last year to $ 69.6 billion. In the Bloomberg survey, net revenue increased by 10 % to $ 24.1 billion, which was 23.5 billion US dollars earlier than the average estimated calculation value.
Microsoft spent $ 22.6 billion in capital expenditures in the second quarter, which was over the previous year. Earlier this month, the company stated that it would spend about 80 billion US dollars in the fiscal year as of June 30 to establish a data center infrastructure required for training AI models and deploying applications.
The expenditure of the bumper is because of worrying that China's Deepseek is in the vortex of Wall Street, claiming that it can perform AI tasks and large American companies such as Microsoft Bask Bask Bask Openai, but the cost is only a small part.
CEO Satya Nadella said Deepseek showed “real innovation”, including matching the performance of the O1 model of Openai, and the performance of the O1 model of OpenAI was touted because it could answer to answer More advanced and scientific issues.
“We will see that everyone will be commercialized and will be widely used. Like customers, the beneficiaries of any software cycle are customers.” He said.
Nadella (Nadella) adds one of the courses from transition to cloud computing (when the company processs the computer to the external server outsourcing), as the cost of resource decreases, demand has risen. He added: “Optimization means that it will be more common.”
This software giant has always been one of the main beneficiaries of AI prosperity. The demand for its cloud services has soared, and the enthusiasm of the $ 13 billion partnership with OPENAI has promoted its results over 3TN.
The OpenAI model is based on Microsoft's Copilot Chat Bot and its large language models, which can be used through Azure.
The group said last week that it will change the transaction structure with Openai so that startups can use the competitors' cloud computing services. But it reserves the right to reject the first time.
This is because the startup company and the cloud provider Oracle and Japan's SoftBank Formulation Plan are the construction of at least $ 10 billion in AI infrastructure in the United States, which is part of the project called Stargate. Microsoft is only appointed as a technical partner of the project.
Nadella said: “Their success is our success because all other business arrangements,” said that before he warned the company to optimize the cost, he referred to the profit distribution arrangements and applications and intellectual property of Microsoft and Openai. “You don't want to buy anything at a time.”
The company added that Microsoft's commercial demand in the quarter increased by 67 %, which was largely driven by OpenAI's commitment.