Ken Griffin calls AI a potential 20-30 year story — not a 3-5 year one — echoing the .com arc
Oct 7, 2025
Key Points
- Ken Griffin frames AI as a 20-to-30-year cycle rather than a 3-to-5-year bet, drawing parallels to the dot-com arc where the Internet's transformative power took far longer to materialize than markets expected.
- Griffin's credibility on this call stems partly from Citadel's lack of exposure to AI labs, unlike competitors like Jane Street, positioning him as a sidelined observer rather than a conflicted stakeholder.
- Markets are pricing near-term AI returns, but most AI-heavy companies remain unleveraged, leaving room for sustained capital intensity before any potential reckoning.
Summary
Ken Griffin frames AI as a 20-to-30-year story, not a 3-to-5-year one, drawing explicit parallels to the dot-com cycle. Speaking at the Future of Global Markets conference, the Citadel CEO argues that language models may take far longer to deliver on their promises than current market sentiment suggests. No one doubted the Internet would change the world, but it took longer than anticipated and involved a real sorting of winners and losers. Griffin expects AI to follow a similar arc.
Griffin's framing carries particular weight because Citadel has no major financial stake in AI labs, unlike Jane Street, which has large positions in Anthropic. The absence of skin in the game creates its own conflict of interest: Griffin is incentivized to be bearish on companies his rivals have bet on.
The broader tension is that markets are pricing in near-term AI returns while Griffin suggests the economic value may diffuse over decades rather than quarters. AI hype has not yet met the combination of low interest rates and leverage that amplified the dot-com bubble. Paul Tudor Jones recently called for a potential "blow off top" before any correction, and most AI-heavy companies are not yet levered, leaving room for further capital intensity before any reckoning.
Calling January 1999 as the parallel is bold because the actual NASDAQ peak came in March 2000, giving investors 15 months of runway to make money before the top. That is plenty of room.
Griffin's credibility rests on being sidelined and intelligent, not on cheerleading. If you are trying to understand where this cycle actually sits, listen to people without massive positions in the game.