Supertake
Illustration for The Electron Bottleneck

The Electron Bottleneck

· 1

−5.2%

since Sep 30

Practice

The idea

I believe AI's real bottleneck isn't chips, it's electricity: turbines, transformers, switchgear, and cooling all have to be built before a GPU ever runs. Rather than betting on one link in that chain, this take spreads across power generation, grid hardware, and inside-the-data-center equipment, blending steadier large caps with higher-torque names. It holds a cash reserve to add on weakness and is built to be held for years, so ordinary industrial-cycle swings don't shake it. I'd rethink it if power actually stopped being the constraint on AI buildout.

What it owns

  • · , , , , , , 32%
  • · , , , , , 27%
  • · , , , 22%
  • · , 11%
  • Cash8%