Apollo’s Torsten Slok warns AI spending isn’t boosting profits outside Big Tech, raising repricing risk

Here’s a number that should make you uncomfortable: the Magnificent Seven tech companies have seen their profit margins climb from roughly 15% in Q1 2023 to around 25% in Q1 2026. The other 493 companies in the S&P 500? They’re still parked at about 10%, exactly where they were three years ago.

Torsten Slok, chief economist and partner at Apollo Global Management, laid out this gap in a Daily Spark note dated June 30, 2026. His core argument is straightforward: if the hundreds of billions flowing into AI infrastructure aren’t actually making non-tech companies more profitable, then the valuations propping up the companies selling AI tools might be built on sand.

The ROI problem nobody wants to talk about

Companies in healthcare, banking, energy, manufacturing, defense, pharma, transportation, construction, and a dozen other sectors are spending on AI but not seeing it show up in their bottom lines. Profit margins across these industries have barely budged.

Slok warned this creates the conditions for what he called a “painful repricing” of AI company valuations. The logic is almost mechanical. If the customers buying AI products aren’t getting richer from using them, eventually they slow down. When that happens, the revenue growth baked into Big Tech stock prices starts looking optimistic.

The economist pointed to several early warning signs already visible in the market: difficulties with model routing, rising token costs, and broader marketplace challenges that suggest implementation is harder than the pitch decks promised.

Why this echoes the dot-com playbook

Slok has drawn comparisons to the dot-com era before, and this latest note reinforces that framing. Capital-intensive and heavily regulated sectors, think utilities, insurance, and defense, face structural barriers to rapid AI adoption. Compliance requirements, legacy systems, union contracts, and safety protocols all add friction. A hospital can’t deploy an AI diagnostic tool the same way a software startup ships a chatbot.

Slok emphasized that sustaining Big Tech’s current valuations requires actual, measurable improvements in economy-wide productivity. Not just more GPU sales. Not just bigger cloud computing bills. Real output gains that justify the trillions in market cap sitting on top of a handful of companies.

What this means for crypto and risk assets

There’s also a funding angle. Many crypto projects, particularly those in the AI-crypto intersection, depend on venture capital flows that are influenced by public market sentiment. If major investment entities like Apollo start signaling caution on AI returns, that skepticism tends to filter down into private markets within a few quarters.

For traders, the signal is clear: watch the earnings reports of non-tech S&P 500 companies over the next few quarters. If profit margins stay flat despite rising AI expenditures, the repricing Slok is warning about becomes increasingly likely.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.