AI’S NEXT BOTTLENECK MAY NOT BE CHIPS.
IT MAY BE POWER.
Think about what is happening:
- Bigger AI models
- More inference
- More data centers
- More GPUs
- More networking
- More cooling
- More electricity
The AI race is rapidly turning into an energy + infrastructure race.
$NVDA can build the world’s fastest GPUs.
But those GPUs still need:
- Electricity
- Data centers
- Cooling
- Grid capacity
- Fiber + optical networking
That’s why I’m watching the entire AI infrastructure stack — not just the obvious AI names.
Semis → Memory → Photonics → NeoCloud → Power → Nuclear → Grid
AI doesn’t stop at the server rack.
It reaches all the way to the substation.
That’s the opportunity most investors are still underestimating.
DYOR.