(Reuters) -Foxconn will collaborate with U.S.-based OpenAI to design and engineer data center racks, components, and other artificial intelligence hardware, the Taiwanese company said on Thursday, as both firms seek to capitalize on booming demand for AI infrastructure.

While the initial agreement does not include purchase commitments or financial obligations, OpenAI will have early access to evaluate the systems and an option to buy them, Foxconn said.

The tie-up will allow Foxconn, the world's largest contract electronics maker, an insight into the growing compute needs of large AI firms and help design products that meet the requirements of advanced large language models.

The Nvidia supplier last week offered a bullish outlook on AI-related demand, saying it would be a big driver of 2026 growth as Foxconn rides the data center boom and benefits from the billions invested by Big Tech firms.

The deal expands OpenAI's involvement in hardware design, having partnered with Broadcom to develop its own custom chip, as it seeks to take a hands-on approach to creating AI systems.

OpenAI CEO Sam Altman has said the startup is committed to spending $1.4 trillion to develop 30 gigawatts of computing resources - enough to roughly power 25 million U.S. homes.

Foxconn will manufacture the data center components, which include cables, power systems, and networking equipment at its U.S. facilities, bolstering supply chains and side-stepping any potential tariffs from the Trump administration.

Separately on Thursday, Foxconn announced a joint venture with Alphabet unit Intrinsic to establish general-purpose robotics and automation across manufacturing facilities to speed-up production.

Initially, the collaboration will cover a range of scenarios across assembly, inspection, machine tending and logistics applications.