Anthropic’s successful initial public offering on Thursday might quiet fears that artificial intelligence is a bubble about to burst. Investors clearly remain willing to wager enormous sums. But even if some of today’s biggest AI companies don’t ultimately justify their valuations — about $2 trillion in Anthropic’s case — the underlying technology will transform the economy.
Federal Reserve Chair Kevin Warsh has been considering what this means for monetary policy. At a conference he convened in San Francisco on Monday, Warsh outlined a range of scenarios, including an AI-driven productivity boom, major job displacement and an investment bust. He described the Fed’s proper posture toward AI as “neither advocacy nor opposition.”
Taming inflation remains Warsh’s most important mission. The minutes from the Fed’s October meeting, released Wednesday, say AI-related demand is straining supply. Some on the committee that sets interest rates see the build-out as a “material source of inflation risk.” Companies continue spending heavily on chips, electricity and data centers. Anthropic alone has committed to nearly $900 billion in computing and infrastructure spending over the next five to eight years.
Companies that harness AI should be able to produce more with fewer resources, discover drugs faster, improve services and create products nobody has imagined yet. Competition will spread the gains. If AI substantially boosts productivity, Warsh thinks the economy can grow faster without generating as much inflation.
The Fed best supports innovation by providing stable prices and a financial system capable of absorbing losses. That requires tolerating inevitable failures. Some AI investments will succeed spectacularly; others will incinerate billions. But that’s the way it goes with capitalism. Experimentation and failure help markets discover what creates value. Falling prices for risky AI investments, perhaps counterintuitively, offer evidence that markets are working.
Failed investment does not necessarily mean wasted investment. America’s 19th-century railroad booms produced bankruptcies alongside tracks that connected the country. The dot-com bubble vaporized fortunes but left behind fiber-optic networks that helped power the internet economy that followed. Something similar could prove true of AI. A data center that proves uneconomic for its original owner does not disappear. Neither do its power connections, chips or fiber. They can change hands at lower prices and become the foundation for the next generation of innovation.
The Fed’s October minutes properly flag the “rapid growth of debt” associated with AI, including “off-balance-sheet leasing commitments and circular financing arrangements,” as a cause for concern. Bank supervisors should pay particular attention to highly leveraged companies and make sure their failures would not threaten the wider financial system.
Anthropic’s blockbuster debut is something to celebrate. Private investors are taking risks on a technology with the potential to make workers more productive, accelerate scientific discovery, create new businesses and raise living standards. The Fed cannot know where AI will lead. Neither can Wall Street. That is why markets, not governments or central bankers, should place the bets.
The post Anthropic’s IPO is worth celebrating. So is the Fed’s caution. appeared first on Washington Post.




