A profile of OpenAI's 'builder-in-chief', Greg Brockman (fortune.com)

OpenAI’s president Greg Brockman has emerged as the company’s “builder‑in‑chief,” leading a massive infrastructure push highlighted by a new multiyear, tens-of-billions deal with AMD to deploy hundreds of thousands of chips across OpenAI’s Stargate data‑center campuses — roughly six gigawatts of compute (about three times Hoover Dam’s output). Brockman, long the operational counterpart to Sam Altman, is driving a program OpenAI says could scale to about 30 gigawatts of compute (with roughly $1.4 trillion cited for deployment) to accelerate AGI research and inference — a bet based on continued scaling laws that predict model gains from larger models, more data, and bespoke AI chips. Technically and strategically, the move shifts OpenAI from model-building to owning the full stack: chips, servers, data centers and operational pipelines for training and inference. That concentration of demand raises near‑term supply, energy and financing implications — straining grids, altering power markets, and prompting creative (and potentially risky) financing such as chipmakers backing loans or taking equity; AMD’s deal even gives OpenAI an option for up to 10% stake. The scale promises huge performance and economic leverage if demand and scaling hold, but also creates political pushback, valuation risks from related‑party transactions, and questions about whether sustained demand will justify trillion‑dollar infrastructure bets.