Apollo Global Management chief economist Torsten Slok published a note on October 24, 2026, with a blunt verdict on the AI boom: outside the tech giants, it has not paid off yet. “Despite massive investments and widespread hype around artificial intelligence, the technology has delivered zero financial payoff to corporate America outside of tech,” Slok wrote. His framing turns the usual AI debate upside down. Instead of asking whether the companies building AI can keep growing, he asks whether the hundreds of companies buying AI are getting anything back.
Slok’s test is simple, and that is what makes it hard to dismiss. He splits the S&P 500 into two groups: the Magnificent 7, Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla, and everybody else. That everybody else, the S&P 493, is where the AI story is supposed to be decided, because those are the companies meant to use AI rather than sell it. On revenue per worker, on profit margins, and on productivity, Slok finds the second group flat.
The numbers behind the claim are stark. Revenue per employee has climbed to roughly $780,000 on an inflation-adjusted basis for the Magnificent 7, while the S&P 493 sits around $525,000, according to Slok’s note. Profit margins tell the same story in a different register: nearly 30% for the tech giants, around 10% for the rest. If AI were lifting the whole economy, you would expect those two lines to at least bend toward each other. They haven’t.
That’s the whole argument in one chart.
Why the AI payoff gap matters now
Slok’s note landed in a week when investors were already questioning the AI trade. On Thursday, October 22, the Nasdaq Composite fell 2.6%, its steepest drop since June, as Meta, Amazon, Nvidia, Tesla, Alphabet, and Microsoft all slid. Meta dropped nearly 7% after raising its projected capital spending for the year to between $135 billion and $145 billion, even as its profit fell short of expectations. Amazon fell 4.6% and Nvidia 4.4%, according to Al Jazeera’s market coverage. Apple was the lone Magnificent 7 stock to rise that day, up 1.2%.
Put those two stories together and the tension is obvious. Investors are paying up for AI infrastructure and AI capability, and at the same time one of Wall Street’s most closely watched economists is saying the returns are showing up in exactly seven companies. Everybody else is spending, but the spending is not visible in the results that shareholders actually get paid on.
Slok has been building this case for months. His October 24 note, titled “No Signs of AI Payoff Outside the Tech Sector,” follows related Apollo notes from June 30 and August 9, 2026. The throughline is consistent: AI adoption is real, AI mentions on earnings calls are everywhere, and AI’s effect on mainstream corporate finances is not.
What the numbers actually show
Start with productivity, because that is where a general-purpose technology is supposed to prove itself. Slok’s productivity data, drawn from the Federal Reserve Bank of San Francisco’s series, shows no clear acceleration in total factor productivity growth despite rapid AI adoption and investment. In plain terms, the economy is not producing measurably more output from the same inputs than it did before the AI spending wave began.
Then look at who is doing the spending. Data from Ramp cited by Apollo shows the top 10% of firms account for 99.5% of model serving and inference spending, and 80.4% of the top 1% of AI spenders are venture-backed companies. That concentration matters. If almost all the AI money flows through a small club of firms, most of them startups burning venture capital, it is no surprise that the broader index looks unchanged. The money hasn’t spread far enough to move the averages.
None of this means companies aren’t using AI. They are, in customer service, coding, marketing copy, and back-office work. Slok’s point is that use is not the same as payoff. A tool can be widely adopted and still fail to raise revenue per worker or widen margins, especially if every competitor adopts the same tool at the same time and the gains get competed away in lower prices.
Why Slok doubts a quick turnaround
Slok sounded skeptical that the gap closes on its own. Turning cheaper intelligence into higher revenue growth and higher profit margins is difficult, he argued, because business results rest on things models don’t grasp. “AI may be a great calculator, but in business, intelligence is mostly judgment,” he wrote. He added that AI “does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, all of which shape reputational risk and ultimately drive revenue growth and profit margins.”
That is a direct challenge to the enterprise AI sales pitch of the last two years. Vendors have sold AI as a cost cutter and a revenue engine in one package. Slok is saying the second half of that pitch is unproven, and the first half may be smaller than advertised once you count what it costs to deploy, secure, and supervise these systems. Recent studies pointing to limited productivity gains so far back him up, though researchers disagree about how long the lag should be. Past general-purpose technologies, from electricity to computers, also took years to show up in productivity statistics.
Critics of Slok’s view have a fair reply: it is early. Companies are still learning how to reorganize work around AI, and the payoff from past technology waves arrived only after firms changed processes, not just tools. But early cuts both ways. If the payoff is still years out, the valuations and spending plans priced for a payoff now look aggressive. That is exactly the question rattling markets this week.
What this means for investors and everyone else
For investors, Slok’s framing gives the AI trade a clear test. Don’t just watch Nvidia’s chip sales or OpenAI’s user numbers. Watch the S&P 493. If revenue per worker and margins for those 493 companies start rising, the AI boom has a second engine. If they stay flat while spending keeps climbing, the boom rests on seven stocks and a promise.
For workers, the message is less abstract. The fear that AI would quickly replace large numbers of jobs has not matched the financial data either. Companies are spending heavily on AI while their revenue per employee barely moves, which suggests they are adding AI on top of existing work rather than replacing that work at scale. That could change, but the numbers Slok cites say it hasn’t yet.
And for the tech industry itself, the note is a warning about who pays next. The Magnificent 7 can fund the AI buildout from their own enormous margins. Their customers cannot, at least not from AI-driven profits they haven’t earned yet. At some point, somebody has to show a return. Slok’s question, stripped of jargon, is simple: if not now, when?
Frequently asked questions
What does Apollo mean by zero AI payoff outside tech?
Apollo chief economist Torsten Slok found no evidence that AI has improved revenue per worker, profit margins, or productivity for the 493 S&P 500 companies outside the Magnificent 7, despite heavy AI spending.
How big is the gap between the Magnificent 7 and the S&P 493?
Revenue per employee is roughly $780,000 for the Magnificent 7 versus about $525,000 for the S&P 493, and profit margins are nearly 30% versus around 10%.
Does this mean AI is a bubble?
Not necessarily. Slok’s point is narrower: the payoff has not shown up in mainstream corporate results yet. The technology could still pay off later, but investors cannot point to it in the numbers today.
Why is Slok skeptical the payoff is just delayed?
He argues that turning cheap AI into higher revenue and margins is hard because business results depend on judgment, trust, and accountability, things AI does not understand.
Sources: TechCrunch, Bloomberg, Fortune, Al Jazeera