Apollo economist says AI payoff is missing for most firms

Torsten Slok says margins and revenue per worker remain flat across the S&P 493, casting doubt on whether AI spending is helping non-tech giants.

AI has yet to improve the bottom line for most large companies outside the tech sector, according to a note from Apollo Global Management chief economist Torsten Slok. His argument is direct: despite the rush to invest in artificial intelligence, he sees no clear sign that it has translated into higher profits or better productivity across corporate America beyond tech. That view stands against the broader narrative that AI is already delivering wide business gains. And because it comes from a closely watched market economist, it is drawing attention.

Slok focuses on what he calls the S&P 493 — the S&P 500 excluding Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla. As IT-PUB News notes, his point is that once those major tech names are removed, the picture looks much less convincing for AI’s business impact.

Slok’s S&P 493 analysis strips out the biggest AI winners

Slok’s framework separates the largest technology companies from the rest of the market. That matters because the excluded firms are among the most prominent AI investors and developers, while the remaining 493 companies represent a much broader slice of corporate America.

According to his note, there is no evidence in that group that AI is improving productivity. Profit margins and revenue per worker, Slok said, appear flat. In his view, that means the technology has not yet produced a measurable financial benefit for most large non-tech businesses.

That does not mean AI is absent from those companies. The point, instead, is that spending on AI — and talking about AI — has not so far shown up in the numbers Slok is tracking.

Slok argues AI still falls short on judgment

Slok also questioned whether non-tech companies can easily turn AI into stronger growth. He argued that it will be difficult for corporate America to “turn increasingly cheap intelligence into higher revenue growth and higher profit margins.”

His skepticism goes beyond software costs or the pace of adoption. In Slok’s view, business performance depends on more than processing information quickly. He wrote that “AI may be a great calculator, but in business, intelligence is mostly judgment.”

He pushed that idea further by saying AI does not understand “accountability, trust, culture, unwritten rules, ethical trade-offs or fairness,” factors he said shape reputational risk and, in turn, revenue and margins. His argument is that even if AI can speed up tasks, it may still fall short where human decisions affect how a company is perceived and how it performs.

The AI productivity debate is getting harder to ignore

Slok’s comments land in the middle of a wider debate over whether AI is already changing business performance in a meaningful way. His note says recent studies have also found that AI has yet to deliver significant productivity gains.

That helps explain why the note is resonating. AI has been pitched as a tool that can make companies faster, leaner and more efficient. If those financial effects are still hard to see in broad corporate data, the gap between expectations and results becomes more difficult to dismiss.

For businesses, the practical question is straightforward: are AI projects creating real value, or are they still mostly experimental and expensive? Slok’s answer, at least for now, is that the payoff has not arrived for most firms outside the tech sector.

The note raises a warning for executives and investors

The note does not claim that AI has no use. It argues something narrower, but still significant: broad corporate adoption has not yet translated into higher margins or stronger revenue per worker. For executives, that is a warning against assuming AI spending will automatically improve results.

For investors, the message is similar. If the technology is not yet lifting profitability across most large companies, market expectations may be moving faster than the underlying business impact.

At the same time, Slok leaves room for a longer timeline. He does not say AI cannot become financially meaningful later. He says that, based on the evidence he cites, the promised payoff has not appeared yet in the parts of corporate America outside the tech leaders.

That leaves the current AI debate centered less on whether the technology exists than on when — and whether — it will start changing the economics of ordinary businesses. For now, Slok says the answer is still unclear, and in the numbers he follows, still missing.