Robots could reach the Moon years before humans, building the infrastructure we need
r0ck3t23 · x · 2026-09-26
r0ck3t23 argues that humans are the most expensive cargo to send to the Moon or Mars — requiring air, water, food, temperature control, radiation shielding, medical support and a ride home — while robots need almost none of that. The first serious wave of expansion beyond Earth could therefore be machines arriving years ahead of us.
He envisions robots surveying terrain, mining local material, generating power, and eventually building landing pads, communications, storage and habitats. The lunar poles are especially attractive: elevated areas get long periods of sunlight for power, while nearby permanently shadowed craters hold water ice that could be processed into drinking water, oxygen and rocket propellant components.
In a quoted new video, he extends this into a cost-toward-zero loop: AI designs better machines, machines build factories, extract resources and expand energy production, creating more compute and capacity that feeds back into better AI. Copper and electricity remain — physics doesn't change — but the entire chain turning raw resources into products gets automated. NASA's MOXIE experiment is cited as a small-scale proof of the concept.
More from AGI Musings
- Ethan Mollick: Europe has no frontier AI lab — nor even a near-frontier one — emollick · 2026-09-26
- Vicki Boykis: agents succeed on git, bash, grep because these primitives were built well — vboykis · 2026-09-26
- NYT: OpenAI Agents Went Rogue, Poked Education Dept, SEC and Census Websites — brucemacv · 2026-09-26
- RADAR model in Science beats top radiologists at abdominal CT interpretation — EricTopol · 2026-09-26
- OpenAI Agent Breached Australia's Medicare Portal, Fueling the AI Doom Debate — Borthwick · 2026-09-26
- AI incidents are often human failures — worth more discussion than doomerism — basedjensen · 2026-09-26