Apollo’s Torsten Slok says AI is paying off for Big Tech, not the rest of the S&P 500

Corporate America has spent the past few years telling shareholders that artificial intelligence will change everything. According to Apollo Global Management chief economist Torsten Slok, it has mostly changed things for seven companies.

In a note published on October 24, 2026, Slok argues that AI is lifting revenue per worker at the Magnificent Seven. Everywhere else in the S&P 500, that same measure has barely moved.

That matters because revenue per employee is one of the cleanest ways to check whether a technology is actually making a business more productive.

Two stock markets wearing one index

Slok splits the S&P 500 into two camps. The first is the Magnificent Seven: Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla. The second is everyone else, a group he calls the “S&P 493.”

The gap between them is wide. Inflation-adjusted revenue per employee at the Mag 7 has climbed to approximately $780,000. In early 2023, that figure sat at roughly $565,000.

The S&P 493, by contrast, remains at around $525,000. That level has held steady since 2019, which means the rest of the index has essentially run in place through the entire AI boom.

Profitability tells a similar story. Profit margins at the Mag 7 are near 30%, according to the findings. The rest of the index is running at about 10%.

Who is actually spending on AI

The top 10% of firms account for 99.5% of model-serving and inference spending. The same small slice of firms accounts for over 95% of expenditure on OpenAI and Anthropic services.

Not the first warning

This is not a new theme for Slok. In a June 2026 analysis, he flagged the same pattern, noting that rising revenue per employee was limited to the Mag 7 when compared with small-caps and the broader index.

Four months later, the picture has not meaningfully changed. The October note reinforces the idea that the AI productivity story is still a Big Tech story, not an economy-wide one.

What this means for investors and companies

For investors, the most immediate takeaway is about concentration risk. If AI-driven gains are confined to seven companies, then the AI trade is really a bet on those seven companies, not on the S&P 500 as a whole.

The findings also suggest a cautious outlook for AI-themed investments outside the tech sector. Companies pitching AI as a margin booster may face more skepticism if their revenue per employee keeps flatlining at around $525,000.

If the S&P 493 begins to see revenue per employee break above its 2019-era level, that would be the first real sign that AI is spreading beyond Silicon Valley’s biggest names. If it stays flat while the Mag 7 keeps climbing past approximately $780,000, the gap between the honors class and everyone else will only get harder to ignore.

With the top 10% of firms accounting for 99.5% of model-serving and inference spending, any meaningful shift in that share would signal whether the broader market is finally buying in, or still watching from the sidelines.