AI Summary
5 min readThe AI Super Cycle: Why Gavin Baker Bets on Watts and Wafers
Gavin Baker has been investing in Nvidia for over 20 years, and his current $4.1 billion portfolio is built on a single conviction: AI is not in a bubble, but in a "super cycle" constrained by physical bottlenecks. While much of the market debates whether AI enthusiasm has run ahead of reality, Baker argues the opposite—the real problem is that demand vastly exceeds the world's ability to supply chips, memory, and power. His fund, Atreides Management, focuses almost entirely on the infrastructure layer: the companies that solve constraints in silicon fabrication, memory bandwidth, connectivity, and energy. He holds a large put position against the broader market (QQQ) while going all-in on specific AI infrastructure plays, essentially betting that the AI buildout will thrive even if the rest of the market struggles.
The Bottleneck Thesis: Watts, Wafers, and Tokens
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of Limitless Podcast
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 (00:00) **Introduction: Gavin Baker and the "Super Cycle" Thesis** - Hosts introduce Gavin Baker, a 20-year AI investor who backed Nvidia and Cerebras, and his core argument: AI is not in a bubble but in a super cycle, with returns concentrated in electricity and silicon fabrication, not software.
- 2 (02:28) **Baker's Portfolio: Infrastructure Bottlenecks** - Breakdown of Baker's top holdings, focusing on companies solving physical constraints in AI compute.
- 3 (07:14) **The Barbell Strategy: Old School and Forward-Looking** - Baker combines bets on established infrastructure (Nvidia, Micron) with speculative plays on future AI paradigms.
- 4 (09:58) **The QQQ Hedge: Bearish on Broad Market** - Baker holds a large put position on QQQ (NASDAQ 100 ETF), signaling skepticism about the overall market while bullish on AI infrastructure.
- 5 (11:08) **Four Key Bottlenecks: Where to Invest** - Hosts outline Baker's four identified constraints that define AI investment opportunities.
- 6 (19:16) **Why This Isn't the Dot-Com Bubble** - Baker's counter-argument: current AI spending is not debt-fueled but from cash reserves of top companies (Google, Microsoft, Amazon, Meta).
- 7 (21:36) **TSMC as the Bubble Blocker** - Baker's thesis that TSMC's inability to meet demand (e.g., $2-3T in potential Nvidia sales) keeps growth sustainable.
+ Full timestamped outline available in the app
Guests on this episode
Show Notes
In this episode, we discuss Gavin Baker’s view that AI is a super cycle, with the biggest opportunities in infrastructure, chips, memory, and power rather than software.
It's interesting to pit him against Leopold Aschenbrenner, who has taken a higher octane approach to similar overarching theses.
Gavin's portfolio includes Astera Labs, Cerebras, NVIDIA, Micron, and Unity Software. DYOR!
------
🌌 LIMITLESS HQ ⬇️
NEWSLETTER: https://limitlessft.substack.com/
FOLLOW ON X: https://x.com/LimitlessFT
SPOTIFY: https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQ
APPLE: https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890
RSS FEED: https://limitlessft.substack.com/
------
TIMESTAMPS
0:00 AI Infrastructure Thesis
1:47 Gavin’s Track Record
2:57 Bottlenecks in Chips
5:26 Unity and World Models
7:40 Inference
11:20 Four AI Constraints
17:35 Energy and Space Compute
19:08 This Isn’t Dot-Com
24:10 Supply Constraints
26:06 Conclusion
------
RESOURCES
Josh: https://x.com/JoshKale
Ejaaz: https://x.com/cryptopunk7213
------
Not financial or tax advice. See our investment disclosures here:
https://www.bankless.com/disclosures
More from this podcast
Limitless Podcast →