AI margin expansion

The chart below plots profit margins for three groups since 2015: the Magnificent 7, the S&P 493 (the S&P 500 minus those seven names), and the broader Bloomberg 500. The Mag 7 line has more than doubled, from around 11% margin in 2015 to roughly 25% today. The other two lines? Basically flat. The S&P 493 has crawled from ~8.5% to ~10% margin over a decade. No trend, just noise around a flat line.

Here’s why that matters right now. Every earnings call outside of Big Tech has some version of “we’re investing in AI to drive efficiency and margin expansion.” It’s the story every CFO is telling. But a decade of data, and the sharpest AI investment cycle in that decade, hasn’t moved the needle for the 493 companies that aren’t building the infrastructure themselves. Meanwhile the bill for the bet is real: hyperscaler capex is tracking toward roughly $730–760B in 2026, but the depreciation being recognized against that spend is only around $211B. The other ~$550B is deferred; It lands on income statements as data centers come online through 2027 and 2028. Free cash flow at the Mag 7 has already compressed to its lowest level since 2024 as capex outpaces revenue. The market needs to see margins move for the S&P 493, not just the Mag 7, or the AI valuation story loses its footing. So here’s the honest timeline question: if AI-driven margin expansion for the other 493 companies hasn’t shown up by 2027–2028, right when the depreciation wave hits hyperscaler earnings and forces a “show me the returns” reckoning, how much longer does the story get to run on CapEx growth and narrative alone?

Want to know more? You can register for free at Fund@mental.

#iamfundamental #soyfundamental #wealthmanagement #familyoffice #financialadvisor #financialplanning #policymistake #ratecut #stagflation