Kevin Warsh thinks AI is going to make us richer, and he thinks getting there will hurt. The Fed official’s framing this week, that the road to an AI-driven wealth boom runs through “very bumpy” terrain before arriving somewhere “very good,” is the kind of hedged prediction that sounds like a cop-out until you look at what markets have actually been doing. Then it starts to read less like cautious central-bank speak and more like a plain description of the last eighteen months.
The bumpiness isn’t hypothetical. We just watched chip stocks shed something on the order of two trillion dollars in a selloff, a correction steep enough to remind everyone that the companies building the AI economy are also the most volatile assets in it. Nvidia, which has become the single clearest proxy for AI enthusiasm, is simultaneously pouring money into Europe and watching its sector get repriced by investors who suddenly remember that valuations can go down as well as up. That tension, between genuine infrastructure demand and the market’s tendency to overshoot in both directions, is the whole story.
Here’s what Warsh’s comment gets right that most bullish AI takes miss: the wealth creation and the wealth destruction aren’t separate events. They’re the same event. Capital floods into a new technology, some of it lands on companies that will define the next decade, and a lot of it lands on companies that will be footnotes. The dot-com era produced Amazon and Google and it also produced Pets.com, and from inside 1999 you could not reliably tell which was which. AI is running that same experiment at a faster clock speed, with more money and more leverage.
The productivity story underneath is real, though. Every credible estimate of AI’s economic impact, whether it comes from central bankers, consultancies, or academic economists, points in the same direction: meaningful gains, unevenly distributed, arriving slower than the hype suggests and faster than the skeptics expect. That gap between when the technology works and when the economy feels it is exactly where the bumpiness lives. Firms invest before they see returns. Workers retrain before they know which skills matter. Markets price in a future that keeps getting delayed by a quarter or two.
And the cost side is getting more interesting by the month. Cybersecurity is a good example. AI is genuinely improving how defenders detect and respond to threats, but it’s also giving attackers cheaper, faster tooling. Regulators are already flagging the risk: FERC has identified compliance gaps in the US power sector tied partly to AI tools and third-party control systems. That’s the pattern. The same capability that creates value creates exposure, and the exposure shows up on a different timeline than the value. You get the wealth later and the risk sooner.
There’s a geographic dimension too. Nvidia’s European push, Samsung’s second chip plant in Taylor, Texas, Hong Kong’s bid to build an embodied-AI hub: these are bets that the physical layer of AI, the fabs, the data centers, the robots, will be spread across multiple regions rather than concentrated in one. That’s good for resilience and bad for anyone who assumed the current supply chain would hold. When capital is this mobile, the bumpiness includes whole countries winning and losing rounds they didn’t see coming.
What should ordinary investors and workers take from all this? Probably not that they should try to time the cycle. The people who got rich in previous technology transitions mostly weren’t the ones who sold at the top. They were the ones who stayed invested, kept their skills current, and didn’t panic when a trillion dollars of paper wealth evaporated over a few weeks. Warsh’s point, stripped of the Fed dialect, is that the destination is probably worth the ride, and the ride is probably worse than you’d like.
The honest version is that nobody, including the Fed, knows how long the bumpy stretch lasts. What we can say with some confidence is that the underlying technology keeps improving, the capital keeps flowing, and the corrections keep coming. A wealth boom that arrives through volatility is still a wealth boom. It just doesn’t feel like one on the days when the screen is red.