Stanford CS153 Frontier Systems | Scott Nolan from General Matter on Energy Bottlenecks

Short summary

Stanford's CS153 Frontier Systems course examines energy as the upstream bottleneck limiting AI compute growth since ChatGPT's 2022 breakout. Guest Scott Nolan (General Matter CEO) argues hyperscalers increasingly require nuclear baseload power for reliability and emissions, but identifies uranium enrichment as America's critical infrastructure gap—the U.S. controls under 0.1% of global capacity. General Matter, founded in 2024 with a $900M DOE contract, aims to reshore this capability.

• •Energy is the upstream constraint on AI's compute scaling, intensified by ChatGPT breakout and enterprise demand
• •U.S. uranium enrichment capacity <0.1% globally; hyperscalers turning to nuclear for baseload power
• •General Matter ($900M DOE contract) founded to reshore American enrichment capacity to Kentucky

Generated with AI, which can make mistakes.