Warsh Channels Greenspan — Fed's AI Productivity Bet
New Fed Chair Kevin Warsh name-checks Greenspan in his first speech, signaling tolerance for above-2% inflation if AI productivity prints follow. What it means for Treasuries, the dollar and AI mega-caps.
Not Just a Tribute — a Policy Tell
Kevin Warsh was sworn in Friday in the East Room of the White House — the first Fed Chair since Alan Greenspan in 1987 to skip the Eccles Building ceremony for a Pennsylvania Avenue setting. He didn't bury the symbolism. In his first speech as Chair, Warsh said he intends to fill the role "with energy and purpose, just the way Chairman Greenspan did," crediting Greenspan as the first person to teach him what the job actually demands.
For desks, that line matters more than the venue. Greenspan's defining call came in 1996/97, when he held rates steady even as unemployment dropped through every NAIRU estimate then in circulation. He read the productivity tea leaves correctly: the PC build-out was lifting trend growth, so the Phillips Curve scream signal was wrong. Anyone short risk on that argument bled out.
The AI Productivity Bet, Operationalized
Warsh has telegraphed the same framework for AI in pre-nomination speeches. Operationally that means a Fed willing to tolerate above-2% core PCE for longer, provided measured productivity is accelerating in parallel. Hikes don't become the default response to sticky services inflation if the productivity print can be argued to absorb it.
Two trades line up cleanly:
- Front-end curve: the market still prices roughly 60 bps of hike risk over the next 12 months. A consistent Greenspan playbook crushes that. 2-year yields have room lower.
- AI capex complex: Nvidia, Broadcom, the hyperscalers — the whole block leans on the assumption that the Fed won't choke the capex cycle. A productivity-friendly Fed extends the runway structurally.
Where the Trade Breaks
It isn't a free option. Greenspan's productivity read was right in 1996. By 2001 it wasn't — the late-90s boom turned into a hard re-rate, and the Fed got blamed for being too patient. If Warsh reads the AI narrative too generously, and the productivity stats don't follow (the AI-ROI debate that pressured names like Microsoft in Q1 hasn't gone away), the setup risks looking 1970s — expectations unanchored, real yields stuck.
Near-term, the tape will read it as it played: long Treasuries, weaker DXY, AI mega-caps preferred. Medium-term, every BLS productivity revision now becomes a binary event for the Warsh trade.