AI has conquered the headlines, the data center plans, and the corporate budget meetings. But has it improved the businesses paying for it? Apollo Global Management chief economist Torsten Slok argues that the answer outside the tech industry is blunt: AI has delivered “zero financial payoff” to corporate America.
The warning lands as companies pour money into compute, chips, data centers, and grid capacity. Profit margins and revenue per worker are rising for the so-called Magnificent Seven, yet the rest of the S&P 500—the S&P 493—has seen no comparable improvement. That gap now puts the entire AI investment story under pressure.
The Business Payoff Has Yet to Appear
Slok says revenue per worker is rising for the Mag 7 but remains flat for the S&P 493. Profit margins show the same pattern, rising for the seven companies while staying flat across the rest of the index. His conclusion is direct: “There are still no signs of AI boosting productivity growth.”
“The bottom line is that the AI story will be decided in the S&P 493, not the Mag 7,” Slok wrote, “and so far, there is no evidence that revenue growth or margins are improving outside the tech sector.” Most companies investing in AI have yet to see significant productivity returns, making the technology look more like a cost center than a growth driver.
That challenge reaches beyond corporate earnings. Slok says companies will only pay a high price for compute when it produces higher revenue growth, bigger profits, or substantial savings. So far, there is little sign that any of those outcomes exist at scale.
“How much compute and how many data centers are needed depends on the returns firms are able to generate with that compute,” he wrote. “So far, the AI payoff for firms has been slow to materialize, implying a growing reluctance to pay for AI services.”
Goldman Sachs CEO David Solomon offered a similar signal, telling investors that “there’s very little that we do where we need frontier models.” Slok says mainstream corporate America does not need the latest AI model—and does not want to pay for it.
Cheap Intelligence Still Needs Human Judgment
The technology may make intelligence cheaper, but cheaper intelligence does not automatically create stronger businesses. Slok says it will be tough for corporate America to “turn increasingly cheap intelligence into higher revenue growth and higher profit margins.”
His explanation focuses on the gap between calculation and judgment. “AI may be a great calculator, but in business, intelligence is mostly judgment,” Slok wrote. “Judgment depends on context, and the answer varies with which human or which model you ask.”
Business decisions also carry risks that do not fit neatly into a calculation. “AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness,” Slok wrote, adding that these factors shape reputational risk and ultimately drive revenue growth and profit margins.
Recent studies point in the same direction. MIT’s Project NANDA studied more than 300 enterprise AI deployments in 2025 and found that 95% delivered zero measurable P&L impact. Stephen Messer summarized the result this way: “Ninety-five percent of enterprise AI deployments deliver zero measurable P&L impact. The other five percent is where the money is.”
Productivity measurement can take time to capture new technology. The American Economic Journal: Macroeconomics published “The Productivity J-Curve” in 2021, a paper by Brynjolfsson, Rock & Syverson. The US Census Bureau confirmed the J-curve pattern in CES Working Paper numbers 25-27, while US Total Factor Productivity was 15.9% higher than official measures showed by the end of 2017. Even with that context, Slok says today’s business results still show no clear AI lift outside tech.
A Trillion-Dollar Spending Bet Meets Rising Pressure
The scale of the investment makes the missing payoff harder to ignore. Enterprise spending on generative AI ran to $30 to $40 billion in 2025 alone, while hyperscalers poured more than $300 billion into data centers, chips, and grid capacity in 2025. A handful of hyperscalers plan to spend an estimated $800 billion this year, with next year’s capex seen topping $1 trillion.
Slok says the core economic flaw in the AI bull case is that “the marginal cost of producing compute exceeds the marginal benefit it delivers to users,” raising questions about claims of unlimited demand for data centers. He also pointed to a slowdown in token spending growth, saying the demand curve for compute is flat and highly sensitive to price changes.
Investors are watching those costs with less patience. “There’s absolutely no patience right now for higher capex,” a portfolio manager said. An investment analyst offered an even harsher image: “we’re going to end up converting all these data centers into pickleball facilities.”
Debt adds another layer of risk. Meta plans to spend $135 billion-$145 billion this year, and its five-year credit default swaps widened 7.6 basis points to an all-time high of 114.79 basis points. CoreWeave stock has crashed more than 30% this month, while its five-year credit default swaps have shot up to 1,145 basis points; CoreWeave has taken on $50 billion in debt.
Anthropic blew through $80.7 billion in the first half of this year. Mark Habib, S&P’s Global Head for Digital Infrastructure, flagged “interconnected financing structures” between hyperscalers, data center projects, and smaller AI companies. “It is safe to assume that there is more (debt) out there that we are not seeing,” he said.
Slok also warned about circularity in token spending. He said 80% of the top 1% of AI spenders are backed by venture capital, with startups paying compute bills using venture capital rather than organic earnings. AI companies have also faced scrutiny over financing deals where they provide money or guarantee debt for other companies that rank among their top customers.
The Next AI Chapter Will Be Measured in Earnings
Slok’s warning does not say AI can never produce returns. “Maybe that will come in the future. Maybe not,” he wrote. But the next stage of the boom will need to prove that models can move beyond impressive calculations and deliver stronger revenue, profits, savings, or productivity.
For now, the biggest test is not how many models companies build or how many data centers they announce. It is whether the S&P 493 can turn AI spending into measurable business results. Until that happens, the technology’s most important number may be the one Slok keeps seeing: zero.
Based on
- Top Economist Finds That AI Has Provided “Zero Financial Payoff” to Companies Outside Tech — futurism.com
- AI delivers ‘zero financial payoff’ to most companies, Apollo economist says | TechCrunch — techcrunch.com
- Hunting for AI gains outside Magnificent 7 | Semafor — semafor.com
- Here’s the ‘core economic flaw’ in the AI bull case as stock and bond trades unravel | Fortune — fortune.com
- AI is everywhere except in the GDP numbers — here is what triggers the break | Fortune — fortune.com
