Artificial intelligence has delivered “zero financial payoff” to corporate America outside the technology industry. That is the blunt verdict from Torsten Slok, chief economist at Apollo Global Management, in a note published recently and reported by TechCrunch on Sunday.
“Despite massive investments and widespread hype around artificial intelligence, the technology has delivered zero financial payoff to corporate America outside of tech,” Slok said.
Slok’s evidence is a construct he calls the “S&P 493”: the S&P 500 minus Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla. Strip out those seven companies — the ones building and selling the AI boom — and what remains, he finds, shows profit margins and revenue per worker that appear flat, with no evidence that AI is improving productivity.
The framing matters because it separates two different stories that usually get blurred together. The first is that AI is a spectacular business for the companies selling it. The second is that AI is making the rest of the economy more productive. Slok’s numbers say the first story is true and the second, so far, is not.
He is skeptical that the gap will close on its own. It will be tough for corporate America to “turn increasingly cheap intelligence into higher revenue growth and higher profit margins,” he argued — a direct challenge to the assumption that falling AI costs automatically become rising profits for everyone.
“AI may be a great calculator, but in business, intelligence is mostly judgment,” Slok wrote. He added: “AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.”
The obvious objection is timing: perhaps companies outside tech are simply late, and the payoff is still coming. Slok’s skepticism is aimed at exactly that defense. If the gains were merely delayed, he implies, the aggregate numbers should at least be bending by now, after years of massive investment and widespread adoption. Instead, they look flat.
There is a caveat the article’s own framing requires, and it should not be buried. Slok’s is one economist’s reading of aggregate indicators, not a measurement of any single company. An index that looks flat can hide real gains at some firms, real losses at others, and long implementation delays in between. And TechCrunch notes that recent studies have also found AI has yet to deliver significant productivity gains — a pattern, not a proof, and one that still leaves room for the payoff to arrive unevenly rather than not at all.
Still, the timing of the note is uncomfortable for the AI spending boom. Amazon CEO Andy Jassy recently said AWS cloud capacity is largely sold out through 2027, with 2028 filling up — demand, in other words, is not the problem. The question Slok raises is what buyers of that capacity get for it. Meta’s Muse assistant passed 10 million users even as AI spending pushed the company’s cash flow negative, a reminder that adoption and profit are different milestones.
What would change the picture? On Slok’s own terms, two numbers: profit margins and revenue per worker for the S&P 493. If AI were paying off broadly, those should rise as deployment spreads. If they stay flat while AI investment keeps climbing, the “zero payoff” claim hardens from provocation into description — and the pressure shifts from AI vendors’ growth stories to their customers’ balance sheets.
One disclosure belongs in any account of the note: Apollo Global Management is the private equity giant that used to own TechCrunch, which reported Slok’s analysis. That history does not invalidate the numbers, but readers weighing a claim about corporate America’s returns deserve to know who published it and through which outlet it reached them.
For now, Slok’s note draws the cleanest line yet through the AI economy: seven companies capturing the upside, 493 still waiting for theirs. Whether that line moves is the number to watch — not the next model release, and not the next funding round, but the margins of the companies buying what the models promise.
