Companies are still shoveling gobs of cash into building out the data centers and other infrastructure for an AI revolution. But voices on Wall Street and elsewhere are once again asking: Is that revolution ever actually coming?
A trio of recent research notes and expert missives have questioned whether the billions of dollars being poured into AI will ultimately pay off. The upshot of these notes seems intended to pour some cold water on runaway hype around data center investment, with warnings that generative AI tech still faces a long road ahead strewn with question marks about its ultimate value.
In a report titled “Gen AI: Too Much Spend, Too Little Benefit?” Goldman Sachs analysts entertained arguments that AI isn’t up to the complex problems it’s been tasked with solving and wondered about its still-TBD “killer application.”
Another colorfully titled research note from Barclays—“Cloud AI Capex: FOMO or Field-Of-Dreams?”—asked whether data center investment is creating a bubble that could end like the telecom crash that followed the 1990s dot-com era. Spoiler alert: “We are leaning FOMO,” the bank’s analysts wrote.
And Sequoia Capital Partner David Cahn recently called the matter of whether the technology can ever recoup massive data center investment “AI’s $600 billion question.”
A lengthy debate: It’s obviously not the first time that investors have asked whether AI spend is headed for a bubble. Questions about the eventual revenue or cost-saving potential of the many AI chatbots and other tools that companies have in the works are a regular feature of any tech earnings call these days.
But many of the strong numbers that have been marshaled to help put this debate to bed—big windfalls from the likes of Microsoft and Nvidia—fall squarely into what Goldman analysts call the “picks and shovels” phase of AI investment: that is, the semiconductors, cloud computing, and energy access that facilitate the creation of AI.
Keep up with the innovative tech transforming business
Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.
On the more skeptical side of the voices included in Goldman’s report is the bank’s head of global equity research, Jim Covello, who expressed doubts that the costs of AI will ever decline enough to make it worthwhile. “To justify those costs, the technology must be able to solve complex problems, which it isn’t designed to do,” Covello said in the note.
Goldman also interviewed MIT economist Daron Acemoglu, who argued that the technology is still far from ready for prime time. “Given the focus and architecture of generative AI technology today, these truly transformative changes won’t happen quickly and few—if any—will likely occur within the next 10 years,” he said.
Dot-com redux? Barclays analysts characterize the thinking of AI boosters as Field of Dreams-style logic, i.e., “If you build it, they will come,” noting that the current projections of capital expenditure is enough to support 12,000 ChatGPT-scale AI products by 2026. Given the amount of spending, Barclays analysts say the industry could be headed to an “overbuild” similar to the telecom crash following the dot-com bubble and “expect someone to flinch next year.”
Deep breaths: Sequoia’s Cahn also noted that AI has a long road ahead of it and warned investors to stay “level-headed” amid “speculative frenzies” like this.
“We need to make sure not to believe in the delusion that has now spread from Silicon Valley to the rest of the country, and indeed the world,” Cahn wrote. “That delusion says that we’re all going to get rich quick, because AGI is coming tomorrow, and we all need to stockpile the only valuable resource, which is GPUs. In reality, the road ahead is going to be a long one.”