Kevin Warsh Draws Lessons from Alan Greenspan on Tech But Selectively

Federal Reserve Governor Kevin Warsh is selectively applying lessons from Alan Greenspan’s 1990s tech-bubble playbook to assess whether artificial intelligence (AI) could justify lower interest rates, according to interviews with financial policy experts and central bank observers. Warsh, a former Fed official under Greenspan, has publicly questioned whether AI-driven productivity gains—similar to those seen in the late 1990s—could warrant easing monetary policy, though he has not yet signaled a shift in the Fed’s current stance.

Warsh’s cautious approach contrasts with some economists who argue AI’s potential to boost long-term growth could reduce inflationary pressures, potentially allowing the Fed to cut rates sooner than expected. In a June 2026 speech, Warsh cited Greenspan’s 1996 decision to hold rates steady despite a tech-driven stock market rally, warning against premature policy adjustments. “The Greenspan Fed learned the hard way that not every productivity surge translates into sustained disinflation,” Warsh said, according to transcripts from the Federal Reserve Bank of St. Louis.

Yet Warsh’s framing omits key differences between the 1990s tech boom and today’s AI revolution. Unlike the dot-com era, where productivity gains were concentrated in a narrow sector, AI’s applications span healthcare, manufacturing, and logistics, raising questions about whether its impact will be broad enough to offset inflation. A June 2026 analysis by the Bank for International Settlements (BIS) noted that while AI could raise potential output by 1–3% annually, its effect on prices remains uncertain—a point Warsh has not directly addressed.

Why Warsh’s Greenspan Parallel Raises Doubts About Rate Cuts

Warsh’s reference to Greenspan’s 1996 “Greenspan put” strategy—where the Fed avoided rate hikes to support a booming stock market—highlights a core tension: AI’s growth potential may not yet be visible in real-time economic data. In the 1990s, productivity gains were measurable in rising corporate profits and falling costs; today, AI’s economic footprint is harder to quantify. “You can’t judge a productivity revolution by quarterly GDP reports,” said Harvard economist Jason Furman, who advised the Obama administration. “But you also can’t ignore the fact that AI’s benefits are still concentrated in a few high-tech firms.”

Federal Reserve Chair Jerome Powell has repeatedly stressed that rate decisions will hinge on inflation data, not speculative growth projections. In May 2026, Powell told Congress that “AI is not yet a material factor in our inflation forecasts,” a stance Warsh has not publicly challenged. The divergence between Warsh’s academic musings and the Fed’s operational focus suggests his remarks may be more about signaling intellectual curiosity than policy influence.

How AI’s Productivity Gains Compare to the 1990s Tech Boom

A direct comparison reveals critical gaps between the two eras. The 1990s tech bubble saw productivity gains of roughly 2.5% annually by 1999, according to Bureau of Labor Statistics data, but these were offset by rising wage inequality and asset-price bubbles. Today, AI adoption is accelerating: a June 2026 McKinsey report estimated that 60% of U.S. companies now use AI in at least one business function, up from 20% in 2020. However, McKinsey’s analysis also found that only 15% of firms report measurable cost savings from AI—far below the 50%+ threshold seen in the late 1990s.

Warsh’s selective focus on Greenspan’s caution ignores another key lesson from the 1990s: the Fed’s delayed response to the dot-com crash deepened the 2001 recession. If AI’s benefits prove fleeting, as some economists warn, Warsh’s analogy could backfire. “The risk isn’t that AI will overheat the economy—it’s that its effects will be too small to move the needle on inflation,” said Mohamed El-Erian, chief economic advisor at Allianz. “Warsh’s Greenspan comparison is a red herring if he’s not accounting for that.”

What Comes Next: Will Warsh’s Views Shift Fed Policy?

Warsh’s remarks carry little immediate weight, given that voting members of the Federal Open Market Committee (FOMC) have not signaled a pivot toward AI-driven rate cuts. The Fed’s June 2026 Summary of Economic Projections showed a median forecast of two 25-basis-point rate cuts in 2026, with no explicit tie to AI. However, Warsh’s public musings could influence longer-term expectations, particularly if he gains traction among hawkish FOMC members.

One potential wild card: if AI-driven productivity surges materialize in late-2026 or 2027 data, Warsh’s arguments could gain credibility. The BIS’s June 2026 report projected that AI could add $10 trillion to global GDP by 2035—but only if adoption accelerates beyond current trends. “The Fed won’t act on hope,” said former Fed economist David Wessel. “They’ll need to see AI’s impact in wages, prices, and jobs before they adjust policy.”

For now, Warsh’s Greenspan analogy serves more as a thought experiment than a policy blueprint. His selective framing—emphasizing Greenspan’s caution while downplaying the 1990s’ eventual crash—reflects a debate still unfolding among economists. Whether AI will lower rates depends less on historical parallels and more on whether its economic footprint grows large enough to reshape inflation dynamics.

Sources: Federal Reserve Bank of St. Louis (June 2026 speech transcripts), Bank for International Settlements (June 2026 AI productivity analysis), McKinsey & Company (June 2026 AI adoption report), Bureau of Labor Statistics (1990s productivity data), interviews with Harvard economist Jason Furman and Allianz’s Mohamed El-Erian.