Apollo Global Management’s chief economist Torsten Slok has put the sceptical case on AI returns in one number: zero. In a Daily Spark note titled No Signs of AI Payoff Outside the Tech Sector, published October 24 and reported by TechCrunch on October 25, Slok splits the S and P 500 into the Magnificent Seven and the remaining 493 companies and finds that profit margins and revenue per worker for the 493 appear flat, with no evidence AI is improving their productivity.
The figures behind the claim are a divergence, not a collapse. Inflation adjusted revenue per employee for the Magnificent Seven has risen from about 550,000 dollars in 2022 to roughly 780,000 to 800,000 dollars in 2026, while the S and P 493 figure has stayed near 525,000 to 550,000 dollars since 2019. Margins divide the same way, near 30 percent for the seven against about 10 percent for the rest. His spending data are more concentrated still: the top 10 percent of firms account for 99.5 percent of model serving and inference spending, 99 percent of neocloud spending and more than 95 percent of spending tied to OpenAI and Anthropic.
Slok’s argument is about demand, not capability. Mainstream companies, he wrote, are starting to view token spending as a cost centre that weighs on margins, most firms do not need the latest model and will not pay for it, and as long as agentic AI stays confined to a small part of the economy its macroeconomic effects will be very limited.
The claim needs its caveats, and they are real ones. Zero is one economist’s reading of aggregate trends, not a company by company audit, and survey evidence cuts the other way in part: McKinsey’s 2026 survey, cited in Dealroom’s account of the note, found 37 percent of respondents attributing at least some earnings impact to AI and 80 percent reporting individual productivity gains, while Deloitte found two thirds reporting efficiency gains but only 20 percent reporting revenue gains. That pattern, workplace gains without enterprise results, is consistent with Slok’s aggregate flatness rather than a refutation of it.