Foxconn, Intel Team Up to Develop AI Infrastructure

As massive AI models scale, especially with the rapid growth of inference and agentic workloads, the need for innovation is accelerating across the full stack of modern computing.

To address data-center scalability challenges, the companies said Thursday that they will focus on developing Intel server racks containing its signature central processing units and advanced AI accelerator architecture. They plan to advance high-speed interconnect technologies, system telemetry and cooling designs.

The alliance pairs Intel’s strengths in processor architecture, silicon technologies and its software ecosystem with Foxconn’s global manufacturing scale and system integration expertise to build integrated, rack-scale AI systems. The companies are designing comprehensive solutions spanning the silicon and module levels up to the rack and system layers, and are also exploring opportunities in custom AI chips.

The partnership extends to edge computing and physical AI, targeting emerging applications such as robotics, automotive, smart cities and smart manufacturing.

In a parallel move Thursday, Foxconn is also deepening its collaboration with Korean conglomerate SK Group to focus on large-scale AI data centers, servers and next-generation memory technologies.

Write to Yang Jie at jie.yang@wsj.com