OpenAI launches US AI supply-chain RFP for manufacturing

This RFP is less about a single procurement cycle and more about codifying OpenAI’s role as a de‑facto industrial policy actor for the AI era. By explicitly targeting U.S. manufacturing of racks, power and cooling gear, robotics components and consumer‑device assemblies, OpenAI is acknowledging that compute scarcity is now as much a physical supply‑chain problem as a cloud‑budget one. In effect, the company is trying to pull an upstream ecosystem of component suppliers into the Stargate gravity well.

For the race to AGI, this matters because it directly attacks one of the few hard bottlenecks that can slow down model scaling: the ability to build and deploy data‑center‑class infrastructure fast enough, in friendly jurisdictions, at acceptable political cost. OpenAI signaling multi‑gigawatt buildout ambitions and then courting hardware partners to match turns infra into a first‑class competitive dimension alongside model quality. It also quietly tilts the playing field toward players who can commit to multi‑year, multi‑billion‑dollar capex in a single country.

The move will likely pressure rivals to deepen their own manufacturing and grid partnerships—whether in the U.S., Europe, or allied Asian hubs—and it will give regulators and communities a clearer target as they weigh the environmental and social costs of AI mega‑projects.