Big Tech Keeps Its A.I. Data Center Spending Boom Alive
![Big Tech Keeps Its A.I. Data Center Spending Boom Alive Big Tech Keeps Its A.I. Data Center Spending Boom Alive](https://i0.wp.com/static01.nyt.com/images/2025/02/08/multimedia/07TECHSPENDING-hqkw/07TECHSPENDING-hqkw-facebookJumbo.jpg?w=780&resize=780,470&ssl=1)
Wall Street went to Panic About two weeks after the start of the Chinese operation Deepseek Issued the artificial intelligence system This seems to be radically more effective than its American competitors.
Investors who have pumped trillion dollars in technology shares over the past few years, whether tens of billions of dollars spent by technology companies spend on new databases suddenly look like Overkill.
But the largest technology companies have shown in the recent profit reports that they believe may not be anything like Overkill when it comes to new data centers.
Amazon implicit Thursday that its capital expenses – the number that includes building the data center and other elements such as warehouses – can reach $ 100 billion this year. Microsoft He said Its spending can exceed $ 80 billion. Al -Abbatit said it would spend $ 75 billion and dead Re -confirmation plans The spending on the capital should reach $ 65 billion.
Completely, you can spend nearly $ 100 billion than last year in these projects.
Executive officials urged patience. They said that the problem at the present time is that customers want more than companies can provide. The only way they can meet the demand is to build as much as possible as possible.
“Whenever I saw another person to do something better, I say,” Laf, we should have done it, “Mark Zuckerberg, CEO of Meta, told the company’s meeting at the company’s level last week, according to a registration obtained from the New York Times. : “The competition is good,” but we need to make sure we win.
Here are some of the main points to understand this moment of technology:
Technology companies need more data centers than they have.
Many companies say they are bound to supply the chips, lands and energy needed to create data centers, and are racing to open more of them. Microsoft, Alphabet and Amazon said they could have had higher cloud computing sales if they had the ability. Cloud services are the typical method that AI is delivered to customers.
Anat Ashkenazi, head of ALPHABET, to Alphabet, told Alphabet, for investors, “The demand” witnessed “the demand that exceeds our available ability.” “So we will work hard to treat it and make sure we are getting more online capacity.”
Microsoft says she is restricting for a while, and has previously told investors that pressure will reduce early this year. But last week, when I reported its latest profits, executive managers told the investors that it might take it until the summer to obtain enough capacity and operate it to meet the full demand. His shares fell about 5 percent in trading category after the report.
They say the most efficiency will expand use and demand for artificial intelligence
While many people think about data centers as expensive places and energy where advanced artificial intelligence systems are developed, they are also where artificial intelligence is spread. These two different steps: a model training that supports ChatGPT, for a chatGPT request for a recipe suggestion.
The spread of artificial intelligence is known as “inference” in the industry; Where technology companies say increasingly, their business will flourish.
“With low costs,” artificial intelligence will be everywhere everywhere. “
Andy Jaci, the Amazon CEO, told investors on Thursday that although the world in which every application has been implanted, it may be difficult for him to understand, “This is the world that we think all the time.” He said this vision has a conclusion in its essence.
He said that reducing the costs of reasoning will follow the pattern of previous technological trends: the systems have become less expensive to publish. “Be enthusiastic about what they could build, as they always thought it was before the cost before, and they usually ended up spending a lot in total. “
Companies say they should think in the long run.
Cloud service providers are used to give customers who are endless supplies, which means that they must reconcile adequate data centers on the Internet for the video flow you want or answer your Chatbot query. But they cannot also build in advance, as they collected billions of dollars that could be published elsewhere. It is one of the tremendous challenges facing companies.
Executive officials have argued that they could adapt how they use investments and build and publish artificial intelligence models, and between their basic business and customer forms. Mr. Nadella said that Microsoft’s infrastructure was “negotiable”. Mrs. Ashkenazi said that Google was also flexible. For example, it can “display the capacity” for Google Search Service instead of cloud customers.
Mr. Zuckerberg said that Meta was studying Deepseek and the methods in which he created efficiency, but investing heavily in data centers will be a strategic advantage against a small and smart competitor.
“We are serving a billion people-this is just a lot of people, so more and more fleet is heading towards inference,” the employees told employees.
Regardless of this interpretation, the intersection is unlikely to lead to profits – even the amazing profits of the largest technology companies – to excite investors. Each company witnessed a decrease in its share price after its profit report.
Niko Grant and Mike Isaac It contributed to the reports from San Francisco.