Overview
- He lays out a five-layer framework for the buildout: energy, chips, physical infrastructure, models, then applications.
- He says the industry is only a few hundred billion dollars into the expansion and that trillions more will be needed, with AI factories rising worldwide.
- He argues the scale-up will generate skilled, well-paid jobs in trades such as electricians, plumbers, pipefitters, steelworkers, and network technicians.
- He distinguishes AI from traditional software by emphasizing real-time, context-driven generation of answers, which he says ties capacity directly to power supply.
- He contends strong open-source models, citing DeepSeek-R1, accelerate adoption and increase demand for training, chips, infrastructure, and energy rather than threatening Nvidia.