Waltstreetbets

The AI arms race is turning Big Tech into glorified utilities and that’s a red flag for investors.

The Magnificent 7 — Meta, Apple, Tesla, Nvidia, Microsoft, Amazon and Nvidia — are being forced to abandon the asset-light business models that fueled their dominance.

In the scramble for AI supremacy, they are becoming asset-heavy industrial firms in a shift that history suggests will erode future returns, according to Kai Wu, founder and chief investment officer of Sparkline Capital.

Featured Video

An Inc.com Featured Presentation

“The irony to this whole thing is that the big tech companies who are actually creating this technology are in many ways actually worsening their position competitively,” Wu told me on Full Signal.

The data justify the concern. Since 2012, the Mag 7’s collective capital expenditures have soared from 4 percent of revenue to 15 percent.

Meta, for instance, plans to spend roughly 35 percent of sales on data centers, a level that matches the average utility and exceeds AT&T’s peak dot-com bubble spending.

Through history, share prices for asset-heavy businesses have underperformed their asset-light peers across all sectors.

What looks like a spending blitz on the surface is masking a fundamental change in market structure.

The economic moats that made these companies distinct — across search, social, retail — are collapsing into a single, viciously competitive arena for general AI.

In Wu’s view, the current boom shares parallels with previous infrastructure buildouts like the 19th century railroads or the internet-era fiber-optic cables, which saw most of the early builders go bankrupt.

The lasting value accrued to the next wave of companies — like Netflix post-dot-com — that leveraged the cheap, overbuilt capacity after the initial boom lost momentum.

This framework suggests the companies most essential to the AI revolution today are the ones actively dismantling the qualities and discipline that made them so critical in the first place.

If the historical pattern holds, the winners of the AI era won’t be the indebted builders of the early years but the companies that arrive later to build transformative businesses on top of the cheap, commoditized capacity left in the boom’s wake.