Seven companies are winning AI. The other 493 earned nothing more

Zero.

That is how much financial payoff artificial intelligence has delivered to corporate America outside of tech, according to the man whose job is to measure exactly that.

Torsten Slok is the chief economist at Apollo Global Management, a firm that manages more than $800 billion. This week he published a note making one blunt point. Strip out 7 companies and the AI boom has not shown up in the numbers at all.

Slok calls the rest of the market the "S&P 493." That is the S&P 500, the list of America's 500 biggest public companies, minus Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla. Those 7 are often called the Magnificent 7, or the Mag 7.

Who is the man saying AI has made no money?

Slok's note matters because he is not a pundit. He studies company numbers for a living, and his view often moves what big investors do with their money.

His point is simple. The AI story everyone talks about, the one about machines doing work and companies getting richer, is really a story about 7 companies. For the other 493, nothing has changed.

How much more are the 7 AI winners making than everyone else?

Slok compared two numbers for each group. Revenue per worker, which shows how much money each employee brings in. And profit margin, which shows how much of that money is kept as profit.

Source: Apollo Global Management, Daily Spark note (24-25 Oct 2026), reported by Yahoo Finance and Bloomberg.

Read nextScammers already know your account details before they call, and that is what makes them so convincing

The 7 AI companies now pull in nearly $800,000 of revenue per worker, up from $550,000 just 4 years ago. Everywhere else in the S&P 500, that number has gone nowhere for 7 straight years.

Margins tell the same story. The Mag 7 now keep almost 30 cents of profit out of every dollar they bring in. The S&P 493 keep about 10 cents. That gap has not closed, even with billions spent on AI tools across nearly every industry.

Those same 7 companies now make up more than a third of the whole S&P 500 index, which is why so much of the market's fate rests on so few names.

Photo: U.S. Department of Agriculture / Wikimedia Commons, Public domain

What does "intelligence is mostly judgment" mean?

Slok put his argument in one line that is already getting passed around trading desks. "AI may be a great calculator, but in business, intelligence is mostly judgment."

His point is that AI is excellent at crunching numbers fast. It is not good at the parts of running a company that actually move profit. He listed what AI cannot do. It does not understand "accountability, trust, culture, unwritten rules, ethical trade-offs or fairness," and he says those are exactly the things that drive revenue growth and profit margins over time.

In other words, a chatbot can write a report. It cannot build the trust that keeps a customer coming back, or make the judgment call that avoids a lawsuit.

Why does my boss say AI is changing everything if the numbers haven't moved?

This is the part that will sound familiar to anyone who has sat through a work meeting this year. Bosses outside tech have spent billions on AI tools, chatbots and automation projects since 2022.

Slok's numbers suggest most of that spending has not shown up as more revenue or fatter profit for the company paying for it. The money and the praise are flowing to the 7 companies selling the AI tools and the chips behind them, not to the companies buying them.

Trending nowPinetown dad fires back at gunmen at his gate, and what the law says

Slok's own words sum it up. "The AI story will be decided in the S&P 493, not the Mag 7." Right now, the 493 have not shown their side of that story yet.

Photo: Farragutful / Wikimedia Commons, CC BY-SA 3.0

What is the $680 billion warning showing up in bond markets?

There is a second worry sitting underneath Slok's numbers. Banking giant Barclays estimates that AI-related debt, money borrowed by chipmakers, data center builders and hardware firms to fund the AI buildout, now adds up to about $680 billion.

That is a lot of debt riding on a bet that AI spending outside the 7 winners eventually pays off. Bond investors are starting to price in doubt about exactly that. Some of the same companies spending the most on AI data centers are the ones whose borrowing costs have started climbing.

What happens this week that could move these numbers?

Three of the Mag 7 report earnings in the next few days. Alphabet reports on 28 October 2026. Microsoft and Amazon follow in the days after. Investors will be watching whether their AI spending is starting to show up as real profit, or just more spending.

The US Federal Reserve also meets on 27 and 28 October 2026 to decide on interest rates, a decision that affects how cheaply all that AI debt can be refinanced.

None of this changes Slok's main number by itself. It takes more than one earnings week to move a gap that has held for seven years. Some firms have already pointed to AI when cutting jobs rather than growing revenue, which is its own kind of answer to where the savings are actually going.

So who is really getting rich off AI?

Picture a gold rush. A handful of people struck gold. Thousands more bought shovels, pans and maps, certain their turn was next.

Right now, 7 companies are the ones who struck gold. The other 493, the airlines, the banks, the retailers, the factories, the hospitals, are still holding the shovel, waiting for their revenue-per-worker number to move even slightly from where it has sat since 2022.