Casey Handmer is the physicist-turned-founder who thinks the entire energy debate is wrong — and he might be right. His company, Terraform Industries, is building shipping-container-sized machines that synthesize natural gas from sunlight and air. His blog has quietly converted thousands of tech-world nuclear enthusiasts into solar bulls. And his January 2026 post "Direct Current Data Centers" dropped a bomb on the AI infrastructure conversation: pure solar + battery can power data centers cheaper than gas, today, with hardware already shipping at prices below the critical threshold. While Microsoft restarts Three Mile Island and Google bets on experimental reactors, Handmer is arguing they're all overcomplicating it.
Dwarkesh is the interviewer who would actually stress-test that claim. His conversations with Dario Amodei, Terence Tao, and Tyler Cowen work because he matches his guests' depth and doesn't let big claims slide. Handmer's thesis — that solar is so cheap the data center is the least interesting thing you can do with it, that the real play is Kardashev Level 1, paving the planet with panels and synthesizing every hydrocarbon from the surplus — deserves exactly that treatment. Not a friendly softball. A real interrogation by someone smart enough to find the holes.
Here's the problem: Dwarkesh has asked his audience for guest suggestions. But suggestions disappear into reply threads. This page doesn't. Every vote here is a permanent, visible signal that this conversation has a real audience waiting for it — one that's harder to ignore than a tweet. If you've read Handmer's blog and thought "someone needs to grill him on this for three hours," this is how you make that happen.
Why Their Voice Would Be Valuable
Ramez focuses on energy, clean tech, and exponential learning curves in solar power that intersect with Casey Handmer's arguments.