Foxconn is investing $690 million in Mexico to establish AI server manufacturing facilities in partnership with Nvidia, with production set to begin in 2025. This represents a new dimension of Mexico's nearshoring story — beyond traditional electronics assembly into the AI compute supply chain. The facilities will assemble GPU server racks, networking equipment, and potentially liquid cooling systems for data centers serving the US market.
The move reflects a broader trend: as US hyperscalers (Microsoft, Google, Amazon, Meta) race to build AI data center capacity, the demand for server hardware is outstripping Asian manufacturing capacity. Mexico's proximity to US data center hubs, combined with USMCA trade advantages and an existing electronics manufacturing workforce, makes it a natural overflow location. The Foxconn-Nvidia partnership specifically targets the assembly of Nvidia's DGX and HGX server platforms.
Strategically, this positions Mexico in the AI hardware supply chain at a critical moment. Global AI server shipments are projected to grow at 30%+ CAGR through 2028, driven by foundation model training and inference demand. If Mexico can establish itself as a reliable nearshore assembly location for AI infrastructure, it captures value from the most capital-intensive technology buildout since the fiber optic boom of the 1990s. The risk is that AI server assembly requires higher precision and cleaner environments than traditional electronics manufacturing.