The Fed just admitted the government might be the bottleneck in the AI productivity boom everyone's banking on.
The Summary
- San Francisco Fed President Mary Daly says regulatory barriers could choke AI's productivity gains, while St. Louis Fed President Alberto Musalem warns AI is currently fueling inflation faster than productivity
- The disconnect: AI is driving demand and investment now, but productivity payoffs are lagging, and fragmented regulations are making it worse for startups
- Translation: The same institution that controls money supply is saying the regulatory state could neutralize the productivity story that's supposed to justify every AI valuation
The Signal
Two Fed presidents just lobbed contradictory warnings about AI within 24 hours. Daly focused on the supply side, arguing that regulatory hurdles create a fragmented landscape that hits startups hardest. If you're building AI tooling or deploying agents at scale, you're navigating state-by-state rules, sector-specific compliance regimes, and federal agencies still figuring out what questions to ask. That friction isn't just annoying. It directly caps how fast productivity can scale.
Musalem took the demand angle. AI infrastructure spending is real. Data center buildouts, GPU procurement, enterprise software contracts—all of it pumps money into the economy today. The productivity gains that are supposed to offset that demand and bring costs down? Still theoretical for most sectors. His message: don't assume AI saves us from inflation. Right now, it's adding to it.
"AI's demand-driven growth fuels inflation, necessitating sustained Fed intervention as productivity gains remain insufficient to lower costs."
Here's what neither Fed official said out loud but both implied: the AI productivity thesis everyone is pricing in might take a decade, not a quarter, to materialize. And in the meantime, regulatory drag and demand-side heat mean the Fed stays cautious. Monetary policy remains "mildly restrictive" according to another Fed voice this week, with inflation risks still "super-elevated." That's Fed-speak for: rates aren't coming down fast, and speculative bets stay expensive.
For builders, this creates a vice. You need capital to navigate the regulatory maze Daly described. But Musalem's inflation warning means the cost of that capital stays high until productivity gains show up in the data. If you're a startup trying to deploy AI agents that automate workflows, you're racing against both the regulatory calendar and the Fed's patience.
The Implication
If the Fed is right, the AI agent economy has a timing problem. Productivity gains are coming, but not fast enough to outrun the demand shock or cut through the regulatory thicket. Companies building automation tools should assume high capital costs persist and plan for longer sales cycles as enterprises wait for clearer compliance paths.
Watch for regulatory fragmentation to favor incumbents who can afford multi-jurisdiction legal teams. Startups without that war chest need to pick their battles: focus on sectors with clearer rules or geographies with unified frameworks. The productivity story is real. The path to it just got longer and more expensive than the market priced in.