OpenAI issues U.S. hardware manufacturing RFP spanning devices, datacenters, and robotics

In January 2026, OpenAI published a document titled "Request for Proposals (RFP) U.S. Hardware Manufacturing," signaling a more explicit push into domestic industrial capacity. In the RFP, OpenAI says it wants to localize significant portions of the manufacturing and assembly behind its hardware devices and AI datacenters over the next 10 years. The stated goals are resilient supply chains, alignment with U.S. values, and stronger national innovation leadership.

The scope is unusually broad. OpenAI is soliciting proposals for U.S.-based manufacturing sites across consumer devices, robotics, and datacenters. For consumer hardware, the list includes final assembly and test, encoders, electronics PCB assembly, advanced node silicon, displays and optics, mechanical tooling, electromechanical modules, manufacturing equipment, packaging, and materials. On the robotics side, OpenAI calls out actuators, precision bearings, harmonic drives, gearboxes, motors, permanent magnets, and power electronics. For datacenters, the document reaches deep into infrastructure, naming backup generators, transformers, UPS, chillers, dry coolers, rear door heat exchangers, coolant distribution units, and cold plates.

The RFP also lays out how proposals will be judged. OpenAI asks respondents to describe location, timeline, manufacturing and technology capabilities, power and utility access, cost structure, and corporate track record. It specifically calls for details on automation, IPC Class 2/3 quality standards, product traceability, MES systems, secure handling of proprietary hardware, and cybersecurity compliance. One notable line is that OpenAI wants meaningful use of automation and AI in the manufacturing process itself, not just in the final product.

The timeline is concrete. Initial proposals are due in June 2026, vendor selection is targeted for March 2027, and joint planning kickoff is scheduled for April 2027. OpenAI says it wants U.S. sites that can come online quickly and expand over time. That suggests the company is thinking beyond model training capacity toward tighter control of the physical systems that may underpin future devices, robotics programs, and data center growth.

The bigger takeaway is strategic. Frontier AI competition is no longer only about models and software APIs; it increasingly depends on power equipment, cooling systems, advanced electronics, and manufacturing execution at industrial scale. By publishing this RFP, OpenAI is making clear that hardware supply chains and factory readiness are becoming core parts of the AI competitive stack.