Tech giants have made clear that they’ll spare no expense in their efforts to win out in the AI rat race. So much so, that tech giants like Meta (META) , Microsoft (MSFT) , Amazon (AMZN) , and Google (GOOGL) planned to spend up to $320 billion on AI tech in 2025.
So when Microsoft CEO Satya Nadella, who augured the great LLM-ification of AI and became a major investor in ChatGPT creator OpenAI in 2019, warned about “overbuild” of data centers after being part of the cohort signaling their spending ambitions, some analysts and industry experts seemed to perk up.
Even more so when Alibaba co-founder Joe Tsai echoed Nadella’s concerns, calling the buildout in AI datacenters a “bubble.” Microsoft, meanwhile, went on to deny additional capacity from hyperscaler CoreWeave (CRWV) , which was in the process of IPOing (that capacity was bought by OpenAI instead.)
Only, nobody really cared.
On Wall Street, valuations in AI-fueled trades were taking off. The only real segway was April’s tariff tiff. Then, it was back to all-time highs for U.S. equities.
The attitude was: “don’t fight the tape.” In some techno-optimist circles, the advent of superintelligent AI was seen just around the corner. Their attitude: “Why sell now?”
Investors might be more wary now, thanks to a recent MIT study warning that businesses are not seeing returns from AI investments. And making matters worse, more industry experts are warning that investors got ahead of themselves.
MIT drops AI spending bombshell
MIT researchers studied 300 businesses and how they were using AI and found that, despite claims that the businesses had invested $30 to $40 billion into generative AI, only 5% of companies had seen any return thus far.
Where industry anecdotes fell on deaf ears, the MIT study cut through the noise. Immediately after the report dropped, so too did tech stocks. And dogpiling on, even more industry leaders are joining the chorus, explicitly calling out what they see as an AI bubble.
OpenAI CEO Sam Altman warns of bubble
Among them are OpenAI CEO Sam Altman, who said in plain terms that, “investors as a whole are overexcited about AI.” While emphasizing its long-term important, Altman cautioned that investors could get “burnt” by the ‘dot com-like’ dynamic in the market.
Unfortunately, it won’t just be the people betting on high-flying names like Palantir (PLTR) . Many Americans’ 401(K)s, IRAs, and brokerage accounts are tied up in indexes which are heavily exposed to the AI trade. In fact, these tech giants represent over a third of the S&P 500’s weight.
Altman remains an optimist over the long run, casting issues with his firm’s latest frontier AI model as a “misfire” and promising an even more fantastic sequel in GPT-6. But that’s what many AI optimists were hoping for GPT-5. And waiting even longer for “the future” to arrive might mean expending their optimism.
That’s not to say that AI models (including at competitors) are not progressing, but what investors’ willingness to allow firms to become capital intensive businesses might not last much longer if they don’t see light at the end of the tunnel.
While they’ve learned to love the stratospheric growth coming from AI chipmaker Nvidia (NVDA) and the double-digit strides in cloud services from Microsoft, Amazon, and Google, wariness about payoff might prompt a pullback.
Mark’s spend-a-thon comes to a close
There’s some evidence that it’s already come — or maybe, somehow, AI is already replacing jobs — at Meta. CEO Mark Zuckerberg spent big to acquire AI talent and build out data centers. He’s now almost fresh out of cash and looking to private credit to shore up his ambitions.
Still, if you’re confident there’s a payoff, why pullback? Per WSJ, Meta is in the process of reorganizing its AI segment into different businesses prioritizing business endpoints. With it, exec departures, layoffs, and a hiring freeze.
Is this a sign that Zuckerberg and management have looked around and collectively discovered that they’re buying the top? Is this an unfortunate repeat of the company’s failed metaverse ambitions? Or is this a wake-up call from within after squandering billions on comp packages for researchers and data centers?
Too early to say, but after blowing through $31.8 billion in the last six months, you’d have to wonder if maybe, the industry gurus called it how it was. Now that Wall Street finally seems to be paying attention, what does that bode for the market?
