Here’s a fun fact that’ll keep you up at night: Amazon, Google, Microsoft, and Meta are collectively spending more money on AI every single month than the entire Manhattan Project cost. Per month. That’s 12 times the Manhattan Project’s annual budget, every year. And according to AI researcher Gary Marcus, this might be the dumbest financial decision in human history.
Marcus, who’s been warning about AI mania longer than most people have been talking about it, dropped a reality check this week after watching Big Tech’s latest earnings bonanza. His verdict? “Greatest capital misallocation in history.” Ouch.
Here’s the thing: these companies are throwing absolutely ridiculous amounts of money at AI infrastructure, and for what? Mark Zuckerberg keeps saying it’s necessary to avoid missing the superintelligence boat. Fair enough—nobody wants to be left behind when the next big thing happens. But here’s the problem: that “next big thing” hasn’t actually materialized yet. And while we’re all waiting, the bills keep piling up.
The numbers are genuinely wild. Each earnings report since 2022 has seen these four hyperscalers set new records for AI spending. Wall Street has mostly cheered them on, rewarding the companies for their commitment to the future. But lately, some cracks are showing. Investors are starting to ask uncomfortable questions, like: “Hey, where’s the actual return on all this money?”
Marcus’s answer is blunt: “None are making major profits on AI. None has a technical moat; a massive price war is inevitable. And few of their customers are seeing major returns on investment.”
Translation: they’re spending like crazy on something that doesn’t actually make them money yet, there’s no competitive advantage to speak of, and the people paying for these services aren’t getting rich off them either. It’s like everyone agreed to buy the most expensive lottery ticket ever created.
This isn’t Marcus’s first rodeo with AI skepticism. Back in February, when tech stocks were getting hammered, he warned that investors had been sold a “bill of goods” on AI. He predicted the stocks would fall further, and honestly, he wasn’t wrong. This week, reports that OpenAI missed key revenue and user growth targets sent tech shares tumbling again.
The irony is delicious: these companies are spending more money than ever on AI, but the market keeps reminding them that maybe—just maybe—the hype has gotten ahead of reality. It’s like watching someone confidently walk off a cliff while insisting they’ve got a parachute that definitely exists and will definitely work.
Now, this doesn’t mean AI isn’t important or that these companies should stop investing. But there’s a difference between strategic investment and throwing money at a wall to see what sticks. Right now, it’s starting to look a lot like the latter.
The real question for investors: how long will Wall Street keep rewarding companies for spending billions on something that hasn’t proven it can actually make them money? Because if Marcus is right—and his track record suggests he might be—that party’s going to end eventually. And when it does, it’s going to be spectacular.
