Video: Opinion | No, Really -- Is A.I. a Bubble?
- Venture capitalist Paul Kedrosky compares the AI boom to railroads, canals, and fiber—all of which ended in massive overbuild and busts.
- He warns that 50% of U.S. household net worth is now in equities, with most tech value tied directly to AI, making your 401(k) a de facto AI bet.
- Upcoming OpenAI and Anthropic IPOs are reportedly larger than all post-WWII IPOs combined—a 'crazy stat' even he admits sounds absurd.
- The bubble isn't about AI being useless; it's about oversupply of capacity and credit-fueled asset inflation, which historically ends in write-downs and failures.
- Kedrosky’s punchline: You don't need a bubble to look like the last one to be just as fragile.
Brief Summary
In a refreshingly blunt segment on 'The Ezra Klein Show,' venture capitalist Paul Kedrosky delivers a eulogy for rational AI valuations before the corpse is even cold. He points to every major capital-expenditure bonanza in Western history—railroads, electrification, fiber—and notes that all of them ended with half the tracks torn up and a pile of bankruptcies. AI, he argues, is not different because it's useful; railroads were useful too, and they still went bust in glorious fashion.
Kedrosky’s real kicker is the math of your retirement account. With half of household net worth now in equities, most of that in tech, and tech mostly AI, the average S&P 500 index holder is effectively a leveraged AI hedge fund. And with OpenAI and Anthropic poised to IPO at valuations bigger than all post-war IPOs combined, he says the market isn't just late to the party—it's passed out in the gutter wearing a party hat.
Why This Matters
If you own an index fund, this is not an abstract debate about tech policy—it’s a direct threat to your retirement. The argument is that AI’s eventual utility doesn't protect you from the financial carnage of overbuilding. The last time we overbuilt fiber, the fiber still worked; it just didn't return 10x to the people who funded it. You’re holding the bag, not the revolution.
The practical takeaway is not to run for the hills screaming, but to understand that the 'AI boom' is now a credit event, not a technology story. If you're young, volatility is a feature; if you're near retirement, this is a warning to rebalance before the trainwreck, not after. Kedrosky’s point is that bubbles don't announce themselves with a foghorn—they look exactly like prosperity right up until they don't.