Torsten Slok of Apollo Global Management argues that revenue per worker and margins remain flat at 493 of the 500 companies in the S&P 500.
Torsten Slok, partner and chief economist at the investment firm Apollo Global Management, published a note on October 24, 2026, arguing that artificial intelligence (AI) has not delivered a financial return to U.S. companies outside the technology sector. According to the text, which can be read on Apollo's website, the gains are concentrated in the big tech companies.
What the report compares
Slok divides the S&P 500 index, which groups 500 large publicly traded companies in the U.S., into two blocks: the seven big tech companies known as “Magnificent 7” and the other 493 companies.
The note examines revenue per worker, adjusted for inflation starting in the fourth quarter of 2019, and profit margins. In both cases, according to Apollo, the figures rise for the Magnificent 7 and are flat for the other 493. The data come from Bloomberg, Macrobond and Apollo calculations. The note's page does not publish numerical values in the text, only in charts.
A third measure concerns productivity. Slok uses a series from the Federal Reserve Bank of San Francisco that adjusts total factor productivity for capacity utilization, and concludes that, on that measure, productivity shows no improvement attributable to AI.
An invoice that reads like a cost
The economist argues that typical companies are starting to see what they pay to use AI models, billed in tokens (chunks of text that are measured and charged), as a cost that cuts into margins rather than a path to growth. According to the note, venture-backed companies spend more on tokens, while most companies do not need the latest models and are not willing to pay for them. Two of the report's five charts use data from Ramp, a corporate expense platform.
Slok adds that AI does not understand responsibility, trust, culture or unwritten rules on its own, factors that influence reputational risk and, ultimately, revenue and margins. His comparison: AI calculates very well, but in a company the judgment of whoever decides carries more weight.
As long as spending on agentic AI and tokens remains a small part of the economy, he warns, the macroeconomic effect will be limited. His conclusion is that the story of AI “will be decided in the S&P 493” and not among the seven largest.
What the report does not show
It is the analysis of an economist at a private fund, not an official statistic. It compares two groups of companies but does not by itself measure whether each company that buys AI services earns a return. Nor does it prove that the return will not come later: Slok suggests the problem could be fundamental and not just a matter of time, but that is his reading.
What does it mean to you?
- If you work for a company that announces AI tools, the report suggests that, for now, that spending is not reflected in higher profits per employee outside the tech sector.
- If you have savings or a retirement plan such as a 401(k) in funds that track the S&P 500, a significant part of its value depends on the seven largest tech companies, according to the report's own analysis. For an investment decision, consult a licensed financial advisor.
- Announcements of cuts or job changes attributed to AI do not equal proven gains: the report does not measure employment.
