San Francisco, United States - December 21, 2025 - OpenAI has entered into a strategic partnership with Taiwan-based manufacturing giant Foxconn to co-design and industrialize specialized hardware for large-scale AI data centres, marking a significant step in OpenAI’s effort to gain tighter control over the physical infrastructure underpinning its compute-intensive models.
Under the arrangement, Foxconn, formerly Hon Hai Precision Industry, will work with OpenAI on the design of servers, racks, and supporting systems optimized for high-density AI workloads. The hardware will be manufactured primarily at Foxconn facilities in the United States, reflecting a broader industry push to localize critical AI infrastructure supply chains amid geopolitical and capacity constraints.
The partnership does not include an immediate binding purchase commitment from OpenAI. Instead, it allows OpenAI to influence hardware architecture early in the design process and evaluate systems that could later be deployed across its expanding data centre footprint. Foxconn, the world’s largest contract electronics manufacturer and a major producer of AI servers, will handle manufacturing and integration, leveraging its scale and experience in high-volume production.
The collaboration is focused squarely on data centre infrastructure rather than consumer devices. Analysts noted that the effort aims to optimize rack-level design, power delivery, cooling efficiency, and system integration to better support large language models and other AI workloads that demand dense compute and high energy efficiency. As AI clusters grow in size, operators are increasingly redesigning hardware layouts to reduce bottlenecks in power distribution and thermal management.
The move comes as OpenAI accelerates investments across the AI infrastructure stack. In recent months, the company has signaled that access to reliable, scalable compute is becoming as strategically important as model development itself. Partnering directly with a manufacturer allows OpenAI to bypass some of the limitations of off-the-shelf systems designed primarily for general-purpose cloud deployments.
Foxconn has positioned the partnership as part of its broader expansion into AI infrastructure manufacturing. The company is ramping up US-based production of servers and related equipment to meet rising demand from hyperscalers and AI developers, with data centres emerging as one of its fastest-growing business segments.
While neither company disclosed timelines for deployment, the partnership underscores a wider shift in the data centre industry toward vertically integrated design, where AI developers, cloud operators, and manufacturers collaborate closely to tailor infrastructure for increasingly specialized workloads.
For OpenAI, the Foxconn deal adds another layer to its long-term strategy to secure compute capacity as competition for AI-ready data centres intensifies globally.
