Apollo's Chief Economist Says AI Has Paid Off for Big Tech and Almost No One Else

Apollo Global Management's chief economist says the AI boom has so far produced a financial payoff for a handful of technology giants and, by his reading of the data, almost nobody else. In a note dated October 24, Torsten Slok wrote that AI "has delivered zero financial payoff to corporate America outside of tech", a line TechCrunch and Tom's Hardware both picked up on October 25.

What the note actually says

The note, published in Apollo's Daily Spark series under the title "No Signs of AI Payoff Outside the Tech Sector", rests on a comparison between two groups of companies. The first is the Magnificent Seven. TechCrunch defines the other group, the S&P 493, as the S&P 500 minus Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla.

Slok's findings are brief. Revenue per worker is rising for the Magnificent Seven but flat for the S&P 493. Profit margins are rising for the Magnificent Seven but flat for the S&P 493. And he sees no signs of AI boosting productivity growth. His conclusion is that the AI story will be decided among the 493, not the seven.

The note contains charts rather than numbers in its text. They draw on Bloomberg, Macrobond, the San Francisco Fed's productivity data, and an AI index and AI spending concentration series from Ramp, along with Apollo's own work. Because the written note gives no percentages, anyone quoting a specific size of the gap is quoting a chart reading, not a stated figure. Tom's Hardware describes the gap in revenue per employee as widening between 2019 and 2026, but it does not give a number either.

Why the finding is narrower than the headline

"Zero financial payoff" is a striking phrase, but it describes aggregate corporate results, not individual deployments. A flat average across 493 companies can hide firms that are saving real money with AI alongside firms that are spending heavily and gaining nothing. The note does not claim otherwise, and it does not say AI tools are useless.

It is also a statement about measured outcomes in a short window. The note does not try to forecast when the picture could change. What it does is put a data point against the assumption that the money now flowing into AI is already earning its keep outside the companies that sell it, which is why the framing around the Magnificent Seven matters. Those seven are both the biggest buyers and the biggest sellers of AI infrastructure, so gains at their end of the market are not evidence of gains for their customers.

There is a disclosure worth knowing. TechCrunch says Apollo is the private equity firm that used to own TechCrunch. The note is a macro view from a large investment firm, not an independent audit, and readers should weigh it as one analyst's reading of public data.

The token cost problem

Tom's Hardware pulled out a second thread from the note: companies are starting to see spending on AI tokens not as a source of growth but as a cost centre. If that holds, the argument for continued spending shifts from "this will grow revenue" to "this will cut costs", a harder case to make when the tools themselves are metered and priced per use. Tom's adds its own analysis on margin risk, which goes beyond what Apollo wrote, so that part is commentary and not a finding.

Tom's also quotes Slok on the limits of the technology itself, saying AI does not understand accountability, trust, culture or unwritten rules. That is an economist's opinion on where machines fall short, not a measurement, and it sits alongside the numbers rather than being proved by them. Taken together, though, the two themes point the same way: the cheap part of AI adoption is buying access, and the hard part is turning it into output that shows up in a company's income statement.

What it means for India

The note covers US companies, so it says nothing directly about Indian firms. Tom's Hardware does cite a separate survey reported by Business Standard in September, in which PwC India found that 84 percent of surveyed fintech respondents had not yet seen measurable financial results from their AI investments. That is a different sample and method from Apollo's, and it covers one sector, so the two should not be merged into a single claim. Together they do suggest that the question of payback is being asked in India as well.

What to watch next

The next test is corporate earnings. If AI spending is starting to pay off outside big tech, it should appear in margins and revenue per worker among the 493 over the coming quarters, and Slok's charts give a baseline to compare against. Watch also for how many companies begin to report AI cost lines separately, since that would show how much of the spending is now treated as a recurring operating expense.

Published October 26, 2026. Gadgets365 will update this article as more information becomes available.