Corporate America is pouring trillions into artificial intelligence, betting on a productivity revolution. But a top economist at Apollo Global Management is asking a crucial question: where is the proof? His analysis suggests the AI payoff is a myth
for most.
The Zero Payoff Argument
In a widely circulated analysis, Torsten Slok, the chief economist at investment firm Apollo Global Management, has delivered a dose of reality to the AI hype train. His argument is stark: despite massive investment, AI has so far produced 'zero financial payoff' for companies outside of the technology sector. Slok separates the 'Magnificent Seven' tech giants from the rest of the S&P 500. What he finds is a tale of two economies. For the tech titans, metrics like revenue per worker and profit margins are climbing impressively. But for the other 493 companies in the index—representing the bulk of corporate America—those same metrics are flat. The promised boom in productivity and profits simply isn't showing up in the broad economic data.
Data vs. The Narrative
The numbers back up Slok's skepticism. According to his research, revenue per employee at the big tech firms surged from roughly $550,000 in 2022 to nearly $800,000 by 2026. For the rest of the S&P 500, that number remained stagnant at around $550,000. Profit margins for the S&P 493 have also held flat at about 10% since 2022, while margins for the tech giants have continued to expand. This data suggests that while AI is generating enormous revenue for the companies selling it (like chip makers and cloud providers), the companies buying and implementing it are not yet seeing a tangible return on their bottom line. In fact, Slok notes that many mainstream companies are beginning to view AI spending not as a source of growth but as a 'cost center that weighs on margins'.
A Modern Productivity Paradox
This isn't the first time a revolutionary technology has failed to immediately deliver on its economic promise. Economists refer to the 'productivity paradox,' a term famously coined in 1987 by Robert Solow, who said, 'You can see the computer age everywhere but in the productivity statistics'. Many experts believe AI is facing a similar 'J-curve' effect, where it takes significant time and investment before widespread productivity gains materialize. Implementing AI isn't like flipping a switch; it requires changes to workflows, employee training, and organizational culture. Proponents argue that we are still in the early days of a transition as profound as the adoption of electricity or the internet, both of which took decades to fully impact the economy. Some studies show specific, task-level productivity gains of 15% to 50% in areas like customer support and marketing, suggesting the potential is there, even if it's not yet visible at a macro level.
The Human Element of Intelligence
Slok's analysis goes beyond the numbers, pointing to a fundamental mismatch between artificial intelligence and business reality. 'AI may be a great calculator, but in business, intelligence is mostly judgment,' he writes. He argues that critical business drivers like accountability, trust, fairness, and understanding unwritten cultural rules are concepts AI cannot grasp. These elements are central to managing reputational risk and, ultimately, driving longterm revenue and profits. This echoes findings from other studies showing that a vast majority of companies investing in generative AI report no measurable impact on their bottom line, with some reporting that as few as 5% of projects deliver transformative returns. The success of AI seems to hinge less on the technology itself and more on its integration into human workflows to solve specific, important problems.
What This Means for the Future
The debate boils down to whether we are witnessing a speculative bubble or the slow start of a genuine economic transformation. The enormous spending on AI infrastructure—data centers, chips, and models—is undeniable. However, Slok's data shows this spending is highly concentrated, with the top 10% of firms accounting for over 95% of spending on major AI platforms. Many of these top spenders are venture-backed startups, not the established mainstream businesses that form the economy's backbone. For investors and business leaders, the message is one of caution. The AI story, as Slok concludes, 'will be decided in the S&P 493, not the Mag 7'. Until broadbased improvements in revenue and margins appear, skepticism about the immediate corporate payoff from AI is likely to grow.
