AI Job Replacement Constrained by Power Supply
Hesamation · x · 2026-07-12
This post cites an energy engineer to counter the narrative that AI will rapidly replace all jobs:
- Replacing 100 million people requires massive compute: Rough estimates suggest that using agents to replace 100 million jobs would demand compute power equivalent to 100 million GPUs.
- Power is a hard constraint: At roughly 700W per agent, the chips alone would require about 70GW. Factoring in cooling and other system overheads, continuous power demand could surge to 100–130GW. For context, the entire US data center system currently consumes only about 20GW.
- Infrastructure build cycles are long: Building new power plants and grid connections typically takes 5–7 years or more.
- Real-world adoption is slower: He also argues that entry-level jobs will indeed be hit first, but this doesn't mean "all jobs will disappear." The claim of a "total replacement within 18 months" is mostly fear-mongering by AI CEOs aimed at selling products, raising capital, justifying layoffs, and depressing wages.
More from AGI Musings
- AI folks, stop saying "biology" when you mean drug development — KevinKaichuang · 2026-10-26
- Anthropic: junior tech hires fell from 27% to 9% as agent hours hit 6x employee hours — thione · 2026-10-26
- Anthropic's Jascha: you may have only 1–5 years of intellectual contribution before AGI — GabGarrett · 2026-10-26
- Anthropic's S-1 Shows $6.6M Revenue Per Employee as AI Agents Log 6x Human Hours — jacalulu · 2026-10-26
- Ben Recht: The world needs a truly open language model, and it's easier to build than you think — beenwrekt · 2026-10-26
- Which jobs fall to AI first? The rule: any layer whose rules can be written as code — FinanceYF5 · 2026-10-26