Apollo’s Torsten Slok Says AI Has Shown No Payoff Outside Big Tech

Artificial intelligence has delivered “zero financial payoff to corporate America outside of tech”, according to Torsten Slok, chief economist at the private equity firm Apollo Global Management. In a Daily Spark note dated 24 October, Slok says revenue per worker and profit margins are rising for the Magnificent 7 but are flat for the rest of the S&P 500.

The note is one economist’s reading of the largest listed US companies, set out in five charts, rather than a study. “Despite massive investments and widespread hype around artificial intelligence,” he writes, the technology has not paid off for “mainstream corporate America”, which he says is “starting to view token spending not as a source of growth but as a cost center that weighs on margins”.

The S&P 493 against the Magnificent 7

Slok’s comparison is between the Magnificent 7 and what he calls the S&P 493. TechCrunch, which reported the note on 25 October, defines that as the S&P 500 minus Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla. TechCrunch also noted that Apollo used to own it.

Apollo’s first chart tracks inflation-adjusted revenue per employee. Read from the chart, the Magnificent 7 line has climbed to around $780,000 in the latest quarter, while the S&P 493 line sits at about $525,000, close to where it has been since 2022. The second chart shows the Magnificent 7’s profit margin near 30%, against about 10% for the S&P 493. Both S&P 493 lines tick up slightly in the most recent quarters, but nothing like the tech group’s rise.

“The bottom line is that the AI story will be decided in the S&P 493, not the Mag 7,” Slok writes, “and so far, there is no evidence that revenue growth or margins are improving outside the tech sector.”

Where the AI money is going

The fourth and fifth charts use data from Ramp. In the fourth, the top 10% of firms account for 99.5% of spending on model serving and inference and 95.6% of spending on OpenAI and Anthropic. In the fifth, 80.4% of the top 1% of AI spenders are venture-backed, against 13.1% of the median spender.

“VC-backed companies are spending more and more on tokens, but mainstream corporate America does not need the latest model and is unwilling to pay for it,” Slok writes. As long as that spending is “confined to a small part of the economy”, he says, “the macroeconomic effects of AI will be very limited”.

That concentration matters for the forecasts behind today’s build-out. Gartner has warned that more than $3.5 trillion of AI infrastructure spending could face write-downs by 2030 if revenue does not keep up. TSN has also looked at what small businesses actually run and at Anthropic’s own data on junior hiring as agents spread inside a tech firm.

What the official productivity data shows

Slok’s third chart, titled “No signs of AI boosting productivity growth”, plots total factor productivity against output per hour since 1985. The latest Bureau of Labor Statistics release, published on 3 September, broadly supports that reading for the economy as a whole, though it covers the whole nonfarm business sector rather than the S&P 493.

The BLS said labour productivity rose at an annualised 1.4% in the second quarter of 2026 and by 2.2% from a year earlier. Since the fourth quarter of 2019, it has grown at an annualised 2.1%. That is faster than the 1.5% of the 2007–2019 cycle, but the same as the long-term rate of 2.1% since 1947.

Slok also doubts that cheaper models will close the gap. “AI may be a great calculator, but in business, intelligence is mostly judgment,” he writes, adding that AI “does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness”.

The gains are real, but they sit in one corner

Slok’s charts stop short of saying AI is failing. They show that its financial benefits so far show up in the companies that sell it, build on it or spend most heavily on it. That is a narrower claim than the “zero payoff” headline suggests, and a more useful one. Revenue per worker and margins are blunt measures, and the BLS figures cover the whole economy, so neither proves AI is doing nothing inside the S&P 493. What they show is that any gains are not yet large enough to move those numbers. For investors, that makes the next few earnings seasons outside tech the real test of whether AI spending turns into returns.

Cheap intelligence still has to become revenue

The note’s sharpest point is about demand, not technology. If mainstream companies see tokens as a cost, falling model prices help only when firms find work worth automating, and Slok’s argument about judgment and accountability suggests that work is harder to find than the price curve implies. That leaves the heaviest AI use inside tech companies and their venture-backed customers, which is exactly where his charts place the spending. The bigger bets on AI agents spending real money assume adoption well beyond that group, and so far the S&P 493 numbers give little sign of it.

Until revenue per worker and margins start rising for the S&P 493, Slok’s charts will keep saying AI’s payoff stops at the tech sector.

Sources

• Apollo, The Daily Spark: No Signs of AI Payoff Outside the Tech Sector, Torsten Slok, 24 October 2026
• US Bureau of Labor Statistics: Productivity and Costs, Second Quarter 2026, Revised, 3 September 2026
• TechCrunch: AI delivers ‘zero financial payoff’ to most companies, Apollo economist says, 25 October 2026

Related stories

• Gartner Forecasts More Than $3.5 Trillion of AI Infrastructure Spending Could Face Write-Downs by 2030
• The $10 Billion Question: What Happens When AI Agents Start Spending Real Money?