OpenAI and Foxconn join forces to build AI hardware in the US

57

In a move that signals deeper ambitions in AI infrastructure, OpenAI has announced a new collaboration with Taiwanese electronics giant Foxconn. The two companies will work together to build AI data center hardware and server components in the United States, aligning with efforts to secure supply chains and meet growing demand for AI computing power.

The partnership will focus on designing custom servers, chips, and networking gear optimised for large language models like GPT-4 and beyond. Foxconn, known globally for assembling Apple iPhones, will bring its manufacturing muscle and supply chain expertise to support OpenAI’s infrastructure roadmap.

This deal comes at a critical time. As the global race for AI supremacy intensifies, infrastructure has emerged as a bottleneck—especially for companies building and deploying large foundation models. OpenAI has been exploring new ways to reduce latency, cut costs, and increase energy efficiency across its data centers.

The alliance also strengthens OpenAI’s strategy to localise more of its AI compute capacity within the US. That includes plans to work with regional partners and chipmakers to build sovereign infrastructure that can support both public and enterprise use cases.

The announcement follows a string of infrastructure-related developments across the AI ecosystem, including Microsoft’s push for custom silicon, Meta’s new AI datacenter plans, and Google’s expansion of its TPU clusters.

By teaming up with Foxconn, OpenAI gains not only a powerful production partner but also access to decades of expertise in hardware miniaturisation and system integration—both critical for scaling future AI models and edge deployments.