Why It Matters
Nvidia is set to launch its Blackwell GPU architecture for AI workloads, with high-profile customers already placing orders for the new chips. To meet the expected demand, Nvidia is collaborating with Taiwanese manufacturer Foxconn to increase availability in the coming months and beyond.
Foxconn is expanding its operations in Mexico with a new plant dedicated to manufacturing Nvidia’s Blackwell architecture, designed to accelerate AI and high-performance computing tasks.
Benjamin Ting, Foxconn’s Senior VP, announced the construction of the world’s largest GB200 production facility in Guadalajara, Mexico, to address the anticipated demand for AI accelerators.
Foxconn, known for manufacturing Apple products, was chosen by Nvidia as a key partner in bringing the Blackwell chip design to reality. The new plant’s production capacity is said to be significant, according to Foxconn Chairman Young Liu.
Aside from Apple products, Foxconn’s production lines also assemble gaming consoles like PlayStation, Xbox, and Switch, along with Amazon Kindle e-readers.
Foxconn’s strategic diversification plan now focuses on AI, with Chairman Young Liu confirming that the new Mexico facility will produce advanced cooling solutions for the GB200 server infrastructure.
Nvidia has positioned Blackwell as a catalyst for a new industrial revolution, offering technology tailored to accelerate AI workloads, minimize inefficiencies, and optimize energy usage.
The upcoming supercomputer powered by the GB200 NVL72 platform, to be jointly built by Nvidia and Foxconn in Taiwan, aims to deliver 90 exaflops of AI performance by 2026 at the Hon Hai Kaohsiung Supercomputing Center.
This supercomputer, set to be Taiwan’s fastest, will be managed by Foxconn for AI-heavy HPC workloads, supporting various applications such as cancer research, AI model development, and smart city initiatives. It will also facilitate digital twins, robotic automation, urban infrastructure enhancement, and AI-driven city services.
Foxconn’s vice-president and spokesman James Wu described the new AI supercomputer as a significant advancement in AI computing and efficiency, positioning it as one of the world’s most potent systems.
The GB200 NVL72 data center platform features 36 Nvidia Grace (Arm) CPUs, 72 Blackwell GPUs, and an NVLink connection delivering 130TB/s of bandwidth in a rack-based infrastructure.
