Nvidia and Foxconn Automate Complex Assembly for AI Server Components
Nvidia and Foxconn have demonstrated automated manufacturing processes for complex components, including busbars for Nvidia's GB300 AI server, achieving over 95% success rates in assembly.
Technology companies Nvidia and Hon Hai Precision Industry (Foxconn) have publicly showcased advancements in automated manufacturing, focusing on complex assembly tasks traditionally requiring significant manual labor. These processes involve components such as busbars and connectors for Nvidia's GB300 AI server system, aiming to improve stability and efficiency.
The automated systems have demonstrated a success rate exceeding 95% for busbar assembly and approximately 90% to 95% for multi-connector insertions. The target assembly time for busbars is under 124 seconds, with multi-connector insertions aiming for under 72 seconds. The overall production cycle is being optimized to not exceed twice the time taken by skilled human workers.
The GB300 is part of Nvidia's Blackwell Ultra architecture, designed for high-performance AI computing. The NVL72 configuration integrates 72 Blackwell Ultra GPUs and 36 Arm-based Grace CPUs into a single, liquid-cooled platform. Foxconn, a major contract manufacturer and a leading supplier of AI servers, is a key partner in the joint development and production of these systems.
This collaboration addresses critical aspects of AI server infrastructure, including high-speed connectors, power delivery solutions like the GB300 Power Clip, and advanced cooling technologies. Foxconn is also exploring the integration of 800 VDC power architectures to enhance the power and cooling efficiency of next-generation AI servers.