If 2023 was the year of the AI chatbot in the tech industry, 2024 is turning out to be the year of AI plumbing. It might not sound as exciting, but tens of billions of dollars are rapidly being spent on behind-the-scenes technology for the industry’s AI boom.
Companies from Amazon to Meta are revamping their data centers to support artificial intelligence. They are investing in huge new facilities, while even countries like Saudi Arabia are rushing to build supercomputers to manage AI. Almost everyone with a foot in technology or giant sums of money, it seems, embarks on a spending spree that some say could last. years.
Microsoft, Meta and Google parent company Alphabet revealed this week that they together spent more than $32 billion on data centers and other capital expenditures in the first three months of the year alone. The companies have all said on calls with investors that they have no plans to slow down their AI spending.
In the clearest sign of how AI has become a story about building massive technology infrastructure, Meta said Wednesday that it needs to spend billions more on chips and data centers for AI than it which she had previously reported.
“I think it makes sense to go for it, and we’re going to do it,” Mark Zuckerberg, Meta’s chief executive, said on a call with investors.
These staggering expenses reflect an old Silicon Valley parable: those who made the biggest fortunes during the California gold rush were not the miners, but those selling the shovels. Nvidia, whose chip sales have more than tripled over the last year, is arguably the most obvious winner when it comes to AI.
Money pouring into technology to support artificial intelligence is also reminiscent of the spending habits of the dot-com boom of the 1990s. For all the excitement about web browsers and new e-commerce websites, The companies that really made money were software giants like Microsoft and Oracle, chipmaker Intel, and Cisco Systems, which made the equipment that connected these new computer networks together.
But cloud computing has added a new problem: Since most start-ups and even large companies in other industries contract with cloud computing providers to host their networks, the tech industry’s largest companies are now spending a lot of money in the hope of attracting customers.
Google’s capital spending — largely money spent on building and equipping data centers — nearly doubled in the first quarter, the company said. Those of Microsoft increased by 22 percent. Amazon, which will publish its results on Tuesday, should contribute to this growth.
Meta investors were upset with Mr. Zuckerberg, causing his company’s stock price to fall more than 16% after the call. But Mr. Zuckerberg, who was pilloried a few years ago by shareholders for a planned spending spree in augmented and virtual reality, has not apologized for the money his company is pouring into AI . He urged patience, potentially for years.
“Our optimism and ambitions have increased significantly,” he said.
Investors had no problem supporting Microsoft’s spending. Microsoft is the only major tech company to release financial details of its generative AI business, which it says has contributed to more than a fifth of the growth in its cloud computing business. This represents a billion dollars in three months, analysts estimate.
Microsoft said its generative AI business could have been even bigger – if the company had enough data centers to meet demand, underscoring the need to keep building.
Investments in AI create a halo for Microsoft’s core cloud computing offering, Azure, helping it attract new customers. “Azure has become a stopover for virtually anyone doing an AI project,” Satya Nadella, Microsoft’s chief executive, said Thursday.
(The New York Times sued Microsoft and its partner OpenAI in December, alleging copyright infringement over news content related to their AI systems.)
Google said sales in its cloud division were up 28%, including “a growing contribution from AI.”
In a letters To shareholders this month, Andy Jassy, Amazon’s chief executive, said much attention had been paid to AI applications, like ChatGPT, but that the opportunity for more technical efforts, around infrastructure and data, was “gigantic”.
For computing infrastructure, “the key is the chip inside,” he said, emphasizing that cutting costs and extracting more performance from chips is key to Amazon’s efforts to develop its own chips. of AI.
Infrastructure demands generally fall into two categories: First, it’s about building the largest and most advanced models, which is what some AI developers say. could soon exceed a billion dollars for each new round. Executives said the ability to work on developing cutting-edge systems, directly or with partners, was essential to staying at the forefront of AI.
And then there’s what’s called inference, or interrogating models to actually use them. This may involve customers leveraging the systems, such as an insurer using generative AI to summarize a customer complaint, or the companies themselves integrating AI directly into their own products, as Meta recently did by integrating a chatbot assistant in Facebook and Instagram. It’s also expensive.
Data centers take time to build and equip. Chips face supply shortages and costly manufacturing. With such long-term bets, Susan Li, Meta’s chief financial officer, said the company is building with “fungibility.” He wants the flexibility to change the way he uses infrastructure, if the future turns out not to be exactly what he expects.
