Aswath Damodaran, the NYU Stern finance professor a lot of Wall Street just calls the “dean of valuation,” has a way of talking about the AI boom that should make its cheerleaders squirm a little. In a recent interview with NDTV Profit, he made a point I keep coming back to: the sky-high valuations we’re pinning on AI aren’t only a bet on the technology. They’re a bet against the workforce. For the numbers to work, the jobs have to go. And once you hear it framed that way, you can’t really unhear it.
The full conversation runs about 27 minutes, and honestly it’s worth watching end to end rather than taking my word for the highlights:
The number that makes the bull case terrifying
This is the part that stopped me. Pressed on what AI actually means for jobs, Damodaran didn’t hand-wave — he did the arithmetic out loud. “The most optimistic stories about AI, the kinds of stories that lead you to put 10 trillion, 15 trillion, 20 trillion as market for AI, are terrifying stories,” he said. To justify something like the roughly $26 trillion figure he saw referenced in the SpaceX prospectus, he reckoned “one out of every two white-collar workers would have to lose their jobs. Lawyers, consultants, bankers, journalists — where do they go?”
And here’s the line that reframes the whole debate for me: “You don’t want the best-case AI scenarios to play out. Because while they may be great for AI companies, it will be catastrophic for the rest of us.” Sit with that for a second, because it means the optimistic case and the pessimistic case aren’t optimistic and pessimistic for the same people. The bull case for the stock is the bear case for your job. Those are the same sentence. It’s also the flip side of a story we’ve covered from the labs’ angle — the way Anthropic has turned what it refuses to build into its whole pitch — except Damodaran is pricing what happens if the build-out actually succeeds.
A great company is not the same as a great investment
The interview opened on SpaceX, which is what he’s gone viral for lately, and I think his read on it is the most quietly useful thing in the whole segment. He’s glowing about the business and merciless about the price, and he doesn’t see any contradiction there. “If you are creating a futuristic company in a movie — think Avengers, right? Stark Industries — you would create a company like SpaceX,” he said, pointing to a company that puts satellites in orbit “at half the cost of anybody else” and is now beaming internet into places like Venezuela that never had broadband. On that last point he’s not wrong about the momentum, either — Starlink just crossed into profitability with four million subscribers.
But loving the engineering doesn’t make it a buy, and this is where most retail investors get tripped up. “I’ve attached a valuation of 1.3 trillion for a company with 20 billion revenues. That requires a lot of story, a lot of potential, and a lot of stuff to come true,” he said — a figure he lays out in detail on his own Musings on Markets blog, where he revisited the SpaceX numbers after the prospectus dropped. Then the seven words I wish every first-time investor would tattoo somewhere visible: “You can have a great company that’s not a great investment.” People conflate those two constantly, and it costs them.
“It’s become this lazy way of saying markets are high because of AI”
He also took a swing at something I’ll admit the media (me included, some days) is guilty of — crediting or blaming AI for every twitch in the market. “If you took the AI companies out of the market, markets are still trading at multiples of earnings that were higher than they were before AI came on,” he pointed out, noting that over the prior two months “the AI trade has actually turned sour for the most part” while the broader market held its ground anyway. His jab at the commentariat landed, and I felt it: whenever he hears that AI alone is holding the market up, “I say you’ve not dug into the data well enough.”
None of this is AI-skepticism, to be clear. He flatly calls it “transformational” and “big.” His actual argument is subtler and, I think, more important — that a huge market and a profitable one are not the same thing. “This can be a big market where the competition is so intense that companies don’t make much money,” he said. The real unknown was never whether AI is a big deal. It’s whether anybody selling the picks and shovels actually gets to keep the money.
Where he thinks the real risk sits
What I appreciated is that he’s not wagging his finger at people for speculating. If you want to pile into an expensive name, go for it — “my advice is mind your own business, just go back and manage your own money.” The exception, and this is the part worth taking seriously, is leverage. The slice of the AI build-out that actually worries him is the part “funded with debt” — not Nvidia, not the megacaps, but the “small to mid-size companies that are using borrowed money to build AI architectures.” When the bill comes due, he warned, “those companies are going to take some of us down with them.” That’s the sentence I’d underline if I were a policymaker.
And his call on the eventual winners is the contrarian bet I find most persuasive. Just like the dot-com era paid off not for the plumbing companies like Cisco but for Amazon, he thinks the AI money ultimately flows to whoever builds profitable products on top of all this infrastructure — a company that, in his words, “might not even be public” yet. If he’s right, half the names people are frantically buying today are the Ciscos of this cycle, not the Amazons. I suspect he’s right. And it fits the shift we’ve been tracking for a while now, where the real disruption is landing on the workforce faster than most companies have redesigned a single job around it.
