OpenAI and Foxconn Partner to Design and Build AI Data Center Hardware in the U.S.

OpenAI has signed a collaboration agreement with Hon Hai Technology Group (Foxconn) to co-design and prepare U.S. manufacturing for the next generation of AI data center hardware. The deal, announced today, covers multiple generations of data center racks plus critical components including cabling, networking, cooling, and power systems — but includes no purchase commitments or financial obligations from either side.

Instead, OpenAI will share insight into emerging hardware needs across the AI industry to inform Foxconn's design and development work. In return, OpenAI gets early access to evaluate the resulting systems and holds an option to purchase them.

The partnership matters because it extends OpenAI's infrastructure push from chips and gigawatt-scale campuses into the physical supply chain that assembles them. OpenAI frames the agreement as part of an effort to strengthen domestic supply chains and, in the company's words, support "continued American leadership in AI."

Three core efforts

The collaboration is structured around three initiatives:

Designing multiple generations of data center hardware in parallel. OpenAI and Foxconn will co-design, engineer, and develop successive generations of AI data center racks simultaneously, pairing OpenAI's infrastructure roadmap with Foxconn's engineering and manufacturing capacity. The stated goals are bringing new systems online faster and securing capacity for long-term growth.

Strengthening and simplifying the U.S. AI supply chain. The companies will improve rack architecture so it can be manufactured across the U.S., broaden sourcing to include more chipsets and domestic suppliers, and expand localized testing and assembly. OpenAI says this will improve reliability, speed deployment, and build what it calls "a more resilient and scalable American supply chain."

Building critical AI data center components in the U.S. Foxconn will manufacture key data center equipment — cabling, networking, cooling, and power systems — at its U.S. facilities. OpenAI positions this as a way to support rapid construction of high-performance compute infrastructure and ensure economic benefits reach American workers and manufacturers.

What the executives said

"We at Foxconn are thrilled to partner with OpenAI—a pioneer at the forefront of the AI digital age," said Foxconn Chairman Young Liu. "As the world's largest manufacturer of AI data servers, Foxconn is uniquely positioned to support OpenAI's mission with trusted, scalable infrastructure that accelerates innovation and broadens access to transformative AI capabilities for businesses and users worldwide."

"The infrastructure behind advanced AI is a generational opportunity to reindustrialize America," said Sam Altman, CEO of OpenAI. "This partnership is a step toward ensuring the core technologies of the AI era are built here. We believe this work will strengthen U.S. leadership and help ensure the benefits of AI are widely shared."

Context and stakes

The agreement fits a broader pattern. As AI models grow more demanding, they require a new class of purpose-built physical infrastructure — racks, cooling, and power delivery tuned to extreme compute densities. Foxconn, already the world's largest manufacturer of AI data servers, gains design guidance from one of the largest buyers of AI compute. OpenAI gains a pathway to hardware supply that does not depend entirely on overseas assembly.

The timing also aligns with U.S. policy pressure to onshore AI manufacturing. OpenAI explicitly argues that building this infrastructure in the U.S. is essential to strengthening supply chains, a framing that tracks with federal incentives for domestic semiconductor and hardware production.

The absence of purchase commitments keeps the financial stakes modest for now. But the option to buy, combined with co-design work spanning multiple hardware generations, signals that OpenAI intends to shape the U.S. AI hardware supply chain directly rather than simply procure from it — a positioning that could matter as competition for AI infrastructure capacity intensifies.