In a nutshell: Nvidia is expected to launch its Blackwell GPU architecture for AI workloads soon, and several high-profile customers have already placed orders to secure the new chips. To meet anticipated demand, Nvidia is partnering with Taiwanese manufacturer Foxconn to significantly boost availability in the coming months and beyond.
Foxconn is expanding its manufacturing operations in Mexico with a new plant dedicated to Nvidia’s Blackwell architecture, the recently unveiled GPU designed to accelerate AI and high-performance computing workloads.
Foxconn Senior VP Benjamin Ting stated that the company is constructing the largest GB200 production facility in the world, aiming to meet the “massive” demand for AI accelerators in the near future.
The plant will reportedly be built in Guadalajara, a Mexican government source confirmed. As the world’s largest electronics contract manufacturer, Foxconn was selected by Nvidia as a key partner in bringing the Blackwell chip design to fruition. Foxconn Chairman Young Liu mentioned that the production capacity of the new plant will be “extremely large,” though he did not provide specific details.
Foxconn is best known as Apple’s primary manufacturing partner, producing iPhones, iPads, and other devices. Its production lines also assemble other notable products, including PlayStation, Xbox, and Switch gaming consoles, as well as Amazon Kindle e-readers.
Foxconn’s next step in its diversification plan focuses on AI. Chairman Young Liu confirmed that the enhanced manufacturing capabilities at the new Mexico facility will support the production of advanced cooling solutions required for the GB200 server infrastructure.
Nvidia has hailed Blackwell as the “spark” of a new industrial revolution, with technology specifically designed to accelerate AI workloads, reduce inefficiencies and downtimes, and optimize energy consumption.
The GB200 NVL72 platform will also power an upcoming supercomputer that Nvidia and Foxconn plan to build in Taiwan. The Hon Hai Kaohsiung Supercomputing Center project, unveiled during Hon Hai Tech Day earlier this week, aims to bring online a high-performance computing system capable of delivering 90 exaflops of “AI performance” by 2026.
Nvidia said the new supercomputer will be Taiwan’s fastest computing infrastructure, with Foxconn managing AI-heavy HPC workloads. The system will be used for cancer research, AI model development, and smart city applications. Additionally, the project will support digital twins, robotic automation, smart urban infrastructure, and AI-assisted city services.
The new AI supercomputer “is one of the most powerful in the world, representing a significant leap forward in AI computing and efficiency,” Foxconn vice-president and spokesman James Wu said.
The GB200 NVL72 data center platform is a rack-based infrastructure featuring 36 Nvidia Grace (Arm) CPUs, 72 Blackwell GPUs, and an NVLink connection capable of delivering 130TB/s of bandwidth.