TAIPEI (Taiwan News) — Foxconn will showcase its manufacturing technologies for AI data centers and AI computing platforms at this year’s Nvidia GTC, highlighting its integrated end-to-end solutions to meet rising demand for AI equipment, CNA reported.

The four-day event will take place in San Jose and run through March 19. It will focus on key topics such as physical AI, AI factory applications, agentic AI, and high-efficiency inference technologies. Nvidia is also expected to unveil its new data center GPU architecture, Feynman, built using TSMC’s 1.6-nanometer process.

Foxconn said it will showcase its manufacturing capabilities for Nvidia’s Vera Rubin full-rack solution at the event. Vera Rubin, the next-generation supercomputing platform that follows Blackwell, is designed with higher-density AI chips and improved energy efficiency.

Foxconn is a major manufacturing partner for the Vera Rubin platform, providing production from chip modules to rack systems. The company said it has developed precision manufacturing and sensor integration technologies, including cold-plate cooling and leak-detection solutions, to help manage the platform’s high power use and heat dissipation needs.

In AI servers, processors generate most of the heat. Foxconn said its cold plates are designed with high‑conductivity materials to move heat away quickly. The company added that, compared with typical cooling methods, its cold plates can boost heat transfer efficiency by more than 30% while keeping the cooling system stable.

According to Nvidia, the company shipped its first Vera Rubin platform in February and plans to begin mass production in the second half of this year.

Foxconn noted that at its November Tech Day event, it showcased an AI modular data center built with Nvidia’s GB300 NVL72 system, highlighting its integration and manufacturing strengths for AI data centers. The GB300 system also provides strong computing power for the company’s smart manufacturing, EV production, and smart city initiatives.

Foxconn Industrial Internet, a subsidiary of Foxconn, reported a significant revenue increase last year due to rising demand for AI. Its AI server business grew more than threefold compared with 2024. The company attributed the growth to its role in the Nvidia supply chain, providing chip substrates and AI data center racks.

Foxconn added that high-end server production requires high-bandwidth memory, whose rising prices could affect profits.