If 2023 was the year of AI chatbots in the technology industry, 2024 will be the year of AI plumbing. It may not sound very exciting, but tens of billions of dollars are quickly being spent on the technology behind the industry's AI boom.
Companies from Amazon to Meta are revamping their data centers to support artificial intelligence. They are investing in massive new facilities, and even places like Saudi Arabia are racing to build supercomputers to handle AI. It seems like almost everyone with a ton of tech or a ton of money is jumping on board a spending frenzy that some think may last. age.
Alphabet, the parent company of Microsoft, Meta and Google, said this week it spent more than $32 billion combined on data center and other capital costs in the first three months of the year. All of these companies said on investor calls that they have no plans to slow down AI spending.
In the clearest sign yet that AI has become a story about building large-scale technology infrastructure, Meta said Wednesday that billions more must be invested in chips and data centers for AI than it had previously signaled.
“I think it makes sense to move forward with it, and we will do it,” Meta CEO Mark Zuckerberg said on a call with investors.
The conspicuous spending reflects an old Silicon Valley trope. The people who made the greatest fortune in the California Gold Rush were not the miners but the shovel diggers. Without a doubt, Nvidia is the clearest AI winner, with chip sales more than tripling over the past year.
The money being invested in technology to support artificial intelligence is also reminiscent of spending patterns from the dot-com boom of the 1990s. For all the excitement about web browsers and new e-commerce websites, the companies that were making the real money were software giants like Microsoft and Oracle, chipmaker Intel, and Cisco Systems, which made the equipment that connected these new computer networks.
However, cloud computing has added new challenges. Big companies in the technology industry are now spending big money to attract customers, as most startups and even large companies in other industries have contracted with cloud computing providers for network hosting.
Google's capital spending – mainly money spent on building and equipping data centers – nearly doubled in the first quarter, the company said. Microsoft rose 22%. Amazon, which reports earnings on Tuesday, is also expected to contribute to this growth.
Meta's investors were not happy with Mr. Zuckerberg, and his company's stock price fell more than 16% after the call. But Zuckerberg, who was criticized by shareholders just a few years ago for planning huge spending on augmented and virtual reality, has made no apologies for the money his company is investing in AI, potentially urging patience for years to come.
“Our optimism and ambition have grown considerably,” he said.
Investors have had no trouble digesting Microsoft's spending. Microsoft is the only major technology company to report financial details of its generative AI business, which it said contributed more than a fifth of the growth of its cloud computing business. Analysts estimated that the amount reached $1 billion in three months.
Microsoft emphasized the need to keep building, saying its generative AI business would have been even bigger if it had secured enough data center supply to meet demand.
AI investments are creating a halo for Microsoft's core cloud computing offering, Azure, helping it attract new customers. “Azure has become the port of call for almost anyone working on an AI project,” Microsoft CEO Satya Nadella said Thursday.
(The New York Times sued Microsoft and its partner OpenAI in December, alleging copyright infringement in news content related to its AI systems.)
Google said its cloud segment revenue increased 28%, including “increased contribution from AI.”
In a letter to shareholders this month, Amazon CEO Andy Jassy said that while a lot of attention has been paid to AI applications like ChatGPT, the opportunity for more technological efforts around infrastructure and data is “huge.”
When it comes to computing infrastructure, “the key is the chips inside,” he said, emphasizing that lowering costs and squeezing more performance out of chips is central to Amazon's efforts to develop its own AI chips.
Infrastructure needs are generally divided into two buckets: First, it is building the largest state-of-the-art model, and some AI developers say it could soon exceed $1 billion with each new round. Top executives said the ability to develop cutting-edge systems, either directly or with partners, is essential to staying at the forefront of AI.
And then there's the task of inferring or querying the model to actually use it. This could include customers leveraging the system, such as an insurance company using generative AI to summarize customer complaints, or companies themselves deploying AI directly into their own products, as Meta recently did by embedding chatbot assistants in Facebook and Instagram. This may include placing, etc. It's expensive too.
Building and preparing a data center takes time. The chips face short supply and expensive manufacturing. With these long-term investments, Susan Li, Meta's chief financial officer, said the company is building with “fungibility.” You need to have the leeway to change how you use your infrastructure if the future turns out not to be exactly what you expected.