OpenAI has partnered with Taiwanese tech giant Foxconn to develop and manufacture new AI data-center hardware, marking another major step in OpenAI’s rapid move into large-scale physical infrastructure. The new systems will be produced at Foxconn facilities in the United States, giving OpenAI early access for testing and an option to buy once development is complete.
The announcement comes as global demand for advanced AI infrastructure far exceeds available supply. OpenAI CEO Sam Altman said the shortage of high-performance components will continue growing as AI adoption accelerates. OpenAI has already committed around $1 trillion to infrastructure deals in 2025, including major partnerships with Oracle and the multibillion-dollar Stargate project involving Oracle and Japan’s SoftBank.
Foxconn, officially known as Hon Hai, has been shifting its focus from traditional electronics manufacturing to high-growth AI server production. The company has seen rising profits as investor interest in AI continues to push valuations upward, even as some analysts warn of growing market volatility.
In a separate update, Foxconn also announced a collaboration with Intrinsic, Alphabet’s robotics unit, to build advanced robotics systems for its U.S. operations. Financial details were not shared.
Industry experts say the new agreement strengthens Foxconn’s position in the expanding AI hardware market. Bloomberg Intelligence analyst Steven Tseng noted that Foxconn has significantly grown its U.S. presence amid geopolitical tensions and concerns over the stability of production in Taiwan. Despite higher costs, he said Foxconn’s scale, strong supply chain, and long experience in server manufacturing make it well-positioned to support OpenAI’s infrastructure goals.
