Meta, Alphabet, and Microsoft are watching their returns on AI spending plummet even as capital expenditure climbs. That's the red flag Aswath Damodaran, NYU's valuation guru, is waving at investors right now. The margin per invested dollar has compressed sharply at all three, a remarkable shift for companies already massive in scale.
The Magnificent Seven Nvidia, Microsoft, Alphabet, Amazon, Meta, Apple, Tesla have enough cash flow and balance sheet muscle to weather what comes next. They've burned tens of billions on AI infrastructure and can absorb losses without breaking a sweat. Their debt capacity is the moat here.
Smaller names have no such cushion
Damodaran's real concern lands elsewhere. Smaller, undercapitalized AI firms lack the financial reserves their bigger peers take for granted. One rough quarter, one pivot in investor sentiment, and they're done. The speed at which capital can evaporate in a downturn keeps getting faster. The hedge fund collapse tied to Situational Awareness proved how quickly AI momentum can shift. "So I think when you see a shakeout in the AI space, it's not so much the Mag-7 we should be watching, but the lesser companies," Damodaran told audiences at NYU Stern.
The math no longer works at scale
Marginal returns on invested capital tell the story better than any market narrative. When you track income gained per new dollar of capex, the numbers at Meta, Alphabet, and Microsoft have deteriorated noticeably. That's not just a sentiment issue. It's mechanical. Spend keeps climbing. Earnings growth doesn't match it. The pattern echoes strain already visible in chipmakers after Micron's sharp share drop rattled the entire memory sector.
Not everyone reads the slowdown as a warning, though. Tom Lee counters that widespread doubt about the AI trade actually suggests the cycle still has runway. But Damodaran's stake is different. If hyperscalers can't post earnings that justify their spending, the next generation of Big Tech will be more capital intensive and deliver lower returns. Whether the correction stays confined to niche AI names or spreads wider probably hinges on one thing: whether hyperscaler spending keeps outpacing earnings growth in the quarters ahead.
This material is for informational purposes only and should not be construed as investment advice or a recommendation to buy or sell any security.
