Play AI CapEx via secondary infrastructure
The guest argued that the multi-year AI CapEx cycle is best expressed through secondary infrastructure enablers rather than direct semiconductor designers.
The argument
The speaker highlighted that instead of owning Nvidia, they focus on power generation, uranium for nuclear energy, grid infrastructure, and optical networking. This approach targets areas with valuation support and asymmetric upside as the massive compute build-out strains existing power and transmission systems.
The thesis, stress-tested
✓ What validates it
- ✓Continued high capital expenditure guidance from major hyperscalers
- ✓Accelerating grid connection approvals for data centers
▸ Risks discussed
- ▸Leaning too hard into one theme and over-concentrating exposure
- ▸Public market valuations running too far ahead of near-term fundamentals
Hear it yourself
"If, you know, you're looking at a manager and they're really impressive and they're really exciting, but they launched their hard tech firm in the beginning of twenty twenty three. How do you assess a track record when so much of it has taken place in in periods where the the wins are behind them? That's a good question, Max. Because I I think, like, just chasing returns, you're gonna chase yourself to a a bottom, because, yeah, the market ebbs and flows. I mean, the Nasdaq's having a good year at 26, decent year. But I think semis are up, like, 50, and software is down 30. And if you lean the wrong way on that, you you know you're dead."
11:15
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE