The “Zero Payoff” Bombshell: Apollo’s Chief Economist Punctures the AI Narrative

While Wall Street keeps cheering the AI arms race, Torsten Slok, Chief Economist of Apollo Global Management — a firm overseeing more than $800 billion in assets — has dropped a bombshell: despite unprecedented corporate investment and pervasive hype, AI has delivered “zero financial payoff” to corporate America outside the tech industry. His research note was picked up by more than a dozen major financial outlets within days, making it the most consequential piece of bearish evidence in the ongoing AI bubble debate.

S&P 493: The Forgotten Sample That Tells the Truth

Slok’s analysis focuses on what he calls the “S&P 493” — the S&P 500 minus Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla. The findings are brutal: across these 493 companies, profit margins and revenue per worker are flat, with no evidence that AI is improving productivity. In other words, the enormous wealth AI has created so far is almost entirely captured by the “Mag 7” — the shovel sellers are getting rich while the miners holding those shovels come up empty.

“AI May Be a Great Calculator, but Business Runs on Judgment”

Responding to the optimistic counterargument that non-tech firms are simply lagging and will catch up, Slok is openly skeptical. He warns it will be tough for corporate America to turn “increasingly cheap intelligence” into higher revenue growth and higher profit margins. His verdict is destined to be quoted for years: “AI may be a great calculator, but in business, intelligence is mostly judgment.”

AI, he argues, does not understand accountability, trust, corporate culture, unwritten rules, ethical trade-offs, or fairness — the invisible soft factors that shape reputational risk and ultimately drive revenue growth and margins. That explains why secretarial work, customer service, and coding can be offloaded to AI while corporate income statements remain unmoved.

A $680 Billion Buildout Meets Bond-Market Warning Lights

The timing makes this more alarming. Alongside Slok’s note, market research indicates that the $680 billion AI infrastructure spending wave has already tripped warning signals in the bond market — rating agencies and credit investors are questioning the sustainability of tech giants funding data center expansion with debt. Absent productivity evidence plus mounting debt is the classic late-stage bubble combination.

The token-cost side presents its own enterprise risk. Accenture recently warned business clients of “sticker shock”: autonomous AI agents burn premium-model tokens at scale, driving corporate AI bills up rather than down. Research also shows a nearly 100-fold price gap between top-tier and entry-level model tokens, yet users keep overusing the most expensive models — a cost squeeze that keeps pressuring corporate margins.

The Counterargument: Measuring the Wrong Layer?

Not everyone is convinced. Critics argue Apollo may be “measuring the wrong layer” — the real risk is not corporate returns but “circular token spend”: tech giants buying each other’s compute and cloud services, recycling revenue inside the Mag 7 flywheel. If enterprise procurement pulls back, the flywheel stalls. On this reading, the S&P 493’s zero payoff is precisely the evidence that AI spending is speculative infrastructure rather than a productivity tool.

Conclusion: Bubble or Not, 2027 Will Decide

Slok’s bottom line is not “AI is useless” but that “the AI story will be decided in the S&P 493, not the Mag 7.” If the profit margins and revenue per worker of those 493 companies start climbing over the coming quarters, AI is the next electrification. If they stay flat, the Mag 7’s AI revenue is a left-hand-to-right-hand capital recycling game — and the bond market has already flashed its first yellow light.

For business leaders the action item is clear: tie every AI project to quantifiable P&L metrics (cost savings, revenue contribution) rather than automation illusions outsourced to agents. For investors, tracking S&P 493 productivity data reveals more about this revolution than tracking Nvidia’s earnings ever will.