FILE – Mark Zuckerberg speaks at Georgetown University in Washington, Oct. 17, 2019. (AP Photo/Nick Wass, File)
SEATTLE — Meta CEO Mark Zuckerberg declared 2023 the “year of efficiency.” Like several of its big tech rivals, Meta has cut jobs and put expansion plans on hold.
Then came AI.
Zuckerberg said earlier this year that his company would spend more than $30 billion on new tech infrastructure in 2024. In April, he raised the figure to $35 billion. On Wednesday, he raised it to at least $37 billion. And he said Meta will spend even more next year.
Zuckerberg said he would rather build “too fast than too late” and give his competitors a big head start in the AI race.
The technology industry’s biggest companies made it clear last week that they have no intention of slowing down their huge spending on artificial intelligence, even as investors grow increasingly concerned that the big gains will come even later than expected.
In the last quarter alone, Apple, Amazon, Meta, Microsoft and Google’s parent company Alphabet spent a total of $59 billion on investments, 63 percent more than the previous year and 161 percent more than four years ago. A large part of this sum went into building data centers and equipping them with new computer systems for developing artificial intelligence. Only Apple has not increased its spending drastically because the company itself does not build the most advanced AI systems.
When investors get nervous, they have to learn to manage their nerves. Last week, Alphabet’s stock price fell more than 5% after the company reported a 91% increase in capital spending. But Alphabet CEO Sundar Pichai pleaded for patience.
“These things take time,” he said, and “the risk of underinvestment is dramatically greater than the risk of overinvestment.”
Executives at the biggest technology companies see a once-in-a-lifetime opportunity in the generative AI technology behind popular chatbots like ChatGPT. They believe it can revolutionize everything from the software that runs the complex operations of global companies to research into new drugs.
When ChatGPT launched in late 2022, tech giants were just beginning to cut spending due to the pandemic, but the industry’s brief austerity policy was over when they realized the potential of artificial intelligence.
This new wave of AI is insanely expensive. The systems work with huge amounts of data and require sophisticated computer chips and new data centers to develop the technology and deliver it to customers. While companies make some revenue from their AI work, it hardly brings anything financially.
In recent months, several high-profile technology industry observers, including the head of equity research at Goldman Sachs and a partner at venture capital firm Sequoia Capital, have questioned whether and when AI will ever deliver enough utility to generate the revenue needed to cover its enormous costs. It’s not clear whether AI will have anywhere near the same impact as the internet or mobile phones, Goldman’s Jim Covello wrote in a June report.
“What trillion-dollar problem will AI solve?” he wrote. “Replacing low-wage jobs with hugely expensive technology is essentially the exact opposite of the previous technology transitions I have witnessed in my 30 years of closely following the technology industry.”
Google’s parent company was the first major technology company to announce its results for the months of April to June, and this heightened concerns about excessive spending. Although Alphabet was able to increase its profits by 29 percent, revenue from ads on Google’s own YouTube fell short of expectations. And the massive increase in infrastructure spending – Google spent an average of $145 million per day in the quarter – unsettled investors.
Next up was Microsoft. As OpenAI’s largest investor, the company had a head start on its rivals and had increased its capital spending every quarter since early last year. But on Tuesday, Microsoft had some somewhat unwelcome news: Its cloud computing business, where most of its AI work was done, wasn’t growing as fast as expected.
But rather than being a sign of caution, the miss (about 1 percentage point below expectations) fueled the building frenzy. Executives said Microsoft has more demand for AI than it can serve from its data centers, a problem they expect to persist through the end of the year. That explains why they are building so feverishly.
Satya Nadella, Microsoft’s CEO, said much of the investment was for the acquisition of land and buildings, which they had to secure in advance, but the remaining 60 percent was for the “kit,” the expensive chips and other components of a computer network, as he called it.
Microsoft executives also asked for patience, saying the spending would generate revenue over the long term – “over 15 years and beyond,” said Amy Hood, the company’s chief financial officer. A look at Microsoft’s numbers shows that the company is on track to generate over $5 billion in revenue from generative AI products this year. But that’s still not much for a tech giant that just reported annual revenue of $245 billion.
The next day, Meta increased its spending forecast. Zuckerberg said he was planning the next generation of AI systems and that the next major update to the company’s main AI model would require 10 times more computing power.
Meta is giving away the advanced AI systems it develops, but Zuckerberg still said it’s worth it. “Part of the great thing about AI is that it can be used to improve all of our products in almost every way,” he said.
Amazon told investors it spent more than $30 billion on capital expenditures in the first half of the year and that it would spend more in the rest of the year as well. Executives said they had to strike a balance between building enough capacity to meet demand without exceeding actual needs.
“The reality right now is that while we’re investing significant amounts in AI and in infrastructure, we’d love to have more capacity than we have today,” said Amazon CEO Andy Jassy. “I mean, we have a lot of demand right now.”
That means buying land, building data centers, and all the computers, chips, and equipment that go with it. Amazon executives see all of that spending as a positive. “We’re using it to grow revenue and free cash flow for the next decade and beyond,” said Brian Olsavsky, the company’s chief financial officer.
There are many signs that the boom will continue. In mid-July, Taiwan Semiconductor Manufacturing Co., which makes most of the chips developed by Nvidia and used in AI systems, said those chips would be in short supply by the end of 2025.
Zuckerberg said the potential of AI was “super exciting.”
“That’s why there are all the jokes about all the tech CEOs talking about AI all the time on these quarterly earnings calls,” he says.
This article originally appeared in the New York Times.
Get more business news by subscribing to our Economy Now newsletter.