Apollo Global Management has delivered a blunt verdict on the AI boom: outside Big Tech, it has not paid off yet. In a note published on October 24, the firm's chief economist, Torsten Slok, found that the so-called Magnificent Seven technology giants are seeing revenue per employee and profit margins rise — while the other 493 companies in the S&P 500 are flat on both measures.
The note, titled "No Signs of AI Payoff Outside the Tech Sector," argues that the fate of the AI trade will be decided not by the handful of companies building the technology, but by the hundreds of companies supposed to be using it. So far, Slok wrote, there is no evidence that revenue growth or profit margins have improved outside the tech sector at all.
The S&P 493 problem
The sharpest part of Slok's argument concerns what he calls the S&P 493: the S&P 500 minus its seven mega-cap tech names. Under his analysis, those 493 companies have seen flat revenue per worker and flat profit margins despite their AI investments, while the Magnificent Seven pull away on both. The San Francisco Federal Reserve's productivity series, he added, shows no sign yet that AI has lifted overall productivity growth.
Token bills become a cost, not an engine
Slok also flagged a slowdown in the growth of token spending — the fees customers pay for the chunks of text AI models process. Earlier in the week he described demand for computing as flat and highly sensitive to price: a buyer who walks away the moment the bill rises. With no improvement in profitability to show for it, some companies are now cutting AI spending, and more treat token costs not as a growth engine but as an expense eating into margins.
The spending that continues is strikingly concentrated. Venture-capital-backed companies keep increasing token outlays, and data from spend-management firm Ramp shows AI spending clustering among a small number of venture-backed firms, while mainstream companies see little need to pay top prices for the latest models.
Slok put the deeper problem in a single passage: "In business, intelligence is mostly judgment. Judgment depends on context, and the answer changes depending on which person or which model you ask. AI does not understand accountability, trust, organizational culture, unwritten rules, ethical choices, or fairness." Those factors, he stressed, protect or destroy corporate reputation — and ultimately determine revenue and profits.
Why the verdict matters — and the pushback
The warning lands as the AI investment boom reaches extraordinary scale. On Bloomberg on October 22, Slok described a widening gap between the massive capital spending of the hyperscalers and stagnant productivity across the broader economy — and warned AI valuations face a "painful repricing" if expected returns fail to arrive quickly. The note struck a nerve, topping Hacker News over the weekend with hundreds of comments split between vindication and scepticism.
The sceptics of the sceptics have a point worth hearing. Independent checks of the same Federal Reserve data note US productivity growth roughly doubled after 2022, even if the component that would show machines making work smarter has gone flat over the past four quarters. Previous transformative technologies, from electricity to information technology, also took a decade or more to show up in aggregate productivity numbers.
That is the real question Apollo has posed: the AI boom is no longer waiting on better models, but on the other 493 companies. Until they show AI in their own revenue and margins, the boom remains a concentrated bet — and concentrated bets are exactly what markets eventually reprice.
