Apollo Global Management chief economist Torsten Slok has argued that artificial intelligence has so far delivered no measurable financial payoff to corporate America outside the technology sector. In a recent note, he compared the seven largest technology companies with the rest of the S&P 500, a group he calls the S&P 493, and found profit margins and revenue per worker essentially flat for the wider group.
Slok’s point is not that AI tools are useless. Tools such as ChatGPT, Claude and Gemini have become everyday productivity aids. His argument is that cheaper intelligence alone does not automatically become higher revenue or fatter margins. As he put it, business success also depends on judgment, accountability, trust, culture and reputational risk, none of which a model supplies on its own.
The note lands at an awkward moment for AI vendors. Companies outside tech have spent heavily on AI, yet many are now treating token spending as a cost to control rather than a growth engine. Recent studies have likewise found little evidence of broad productivity gains so far, and a growing share of corporate AI projects are being scaled back or abandoned.
That does not settle the debate. Supporters argue that gains take time, that firms must reorganise around AI before returns appear, and that early benefits may be captured first by vendors rather than buyers. What Slok’s analysis does show is that the burden of proof has shifted: vendors now need to demonstrate returns in revenue and margins, not just in demos and adoption numbers. For related coverage, see our report on Replit shifting its AI spend to OpenAI in a discount deal.
Source: TechCrunch
