If 2023 was the year of AI chatbots for the tech sector, 2024 is proving to be the year of AI. It may not sound that exciting, but tens of billions of dollars are quickly being spent on behind-the-scenes technology for the industry's AI boom.
Companies from Amazon to Meta are renovating their data centers to support artificial intelligence. They are investing in huge new facilities, while places like Saudi Arabia are also racing to build supercomputers capable of handling artificial intelligence. Nearly everyone with a foot in technology or huge amounts of money, it seems, is jumping into a spending frenzy that some believe could last for a long time. years.
Microsoft, Meta and Google's parent company Alphabet revealed this week that they spent more than $32 billion combined on data centers and other capital expenditures in the first three months of the year alone. All the companies said in investor calls that they had no plans to slow spending on AI.
In the clearest sign yet of how artificial intelligence has become a story about building a huge technological infrastructure, Meta said Wednesday that it would have to spend billions more on AI chips and data centers than it had signaled previously.
“I think it makes sense to try, and we will,” Mark Zuckerberg, Meta's chief executive, said on a call with investors.
The eye-popping expenses reflect an old Silicon Valley parable: The people who amassed the biggest fortunes during the California Gold Rush weren't the miners, they were the people who sold shovels. Without a doubt Nvidia, whose chip sales have more than tripled in the past year, is the most obvious winner in AI.
The money invested in technology to support artificial intelligence is also reminiscent of the spending patterns of the dot-com boom of the 1990s. Despite all the excitement around web browsers and new e-commerce sites, the companies really making money were software giants like Microsoft and Oracle, chipmaker Intel, and Cisco Systems, which made the tools that linked the new ones together. computer networks.
But cloud computing has added a new wrinkle: Since most startups and even large companies in other industries contract with cloud computing providers to host their networks, the largest tech companies are spending much in the hope of attracting customers.
Google's capital spending – much of it money spent building and outfitting data centers – nearly doubled in the first quarter, the company said. Microsoft's increased by 22%. Amazon, which reports earnings on Tuesday, is expected to contribute to that growth.
Meta investors were unhappy with Mr. Zuckerberg, who sent his company's stock price down more than 16% after the call. But Mr. Zuckerberg, who only a few years ago was pilloried by shareholders for a planned spending spree in augmented and virtual reality, has made no apologies for the money his company is investing in artificial intelligence. He urged patience, potentially for years.
“Our optimism and ambitions have grown quite a bit,” he said.
Investors had no problem putting up with Microsoft's spending. Microsoft is the only major technology company to report financial details of its generative artificial intelligence business, which it says has contributed to more than a fifth of the growth of its cloud computing business. According to analysts' estimates, this amounted to $1 billion in three months.
Microsoft said its generative AI business could have been even bigger if the company had had enough data center supply to meet demand, underscoring the need to keep building.
Investments in artificial intelligence are creating a halo for Microsoft's core cloud computing offering, Azure, helping it attract new customers. “Azure has become a one-stop shop for pretty much anyone who is doing artificial intelligence projects,” Satya Nadella, Microsoft's chief executive, said Thursday.
(The New York Times sued Microsoft and its partner, OpenAI, in December, alleging copyright infringement of news content related to their artificial intelligence systems.)
Google said sales from its cloud division rose 28%, including “an increasing contribution from artificial intelligence”
In a letter to shareholders this month, Andy Jassy, Amazon's chief executive, said that a lot of attention has been paid to artificial intelligence applications, such as ChatGPT, but that the opportunity for more technical efforts, regarding infrastructure and data, it is “gigantic”.
For computing infrastructure, “the key is the chip inside it,” he said, noting that reducing costs and getting more performance out of chips is central to Amazon's effort to develop its own AI chips.
Infrastructure requests generally fall into two categories: First, building larger, cutting-edge models, which some AI developers say could soon exceed $1 billion per new round. CEOs said being able to work on cutting-edge system development, directly or with partners, is essential to staying at the forefront of artificial intelligence
And then there's what's called inference, which is the questioning of models to actually use them. This may involve customers tapping into systems, such as an insurer using generative AI to summarize a customer complaint, or companies themselves building AI directly into their products, as Meta recently did by embedding an assistant chatbots on Facebook and Instagram. That's expensive too.
Data centers take time to build and equip. Chips face supply shortages and expensive manufacturing costs. With such long-term bets, Susan Li, Meta's chief financial officer, said the company is building with “fungibility.” It wants leeway to change how it uses infrastructure if the future turns out not to be exactly what it expects.