Nvidia and Foxconn Boost AI Assembly Success with Robots

Nvidia Corp. and Hon Hai Precision Industry Co. (Foxconn) have successfully improved the assembly success rates of Nvidia's GB300 NVL72 rack-scale AI platform by leveraging robotic automation. In a detailed report published on Nvidia's technical blog on October 2, the companies highlighted how robots were trained to handle critical assembly steps, significantly enhancing manufacturing stability and efficiency. Success rates for key assembly tasks exceeded 90%, with the busbar assembly achieving over 95% success and multi-connector insertion ranging from 90% to 95%.

The focus of the automation efforts was on assembling the GB300's tester tray, which is essential for verifying the functionality and performance of the computing modules before deployment. Key tasks included the assembly of busbars, which carry high electrical currents, and the insertion of multiple electrical connectors with tight clearances. The automation system aimed to complete busbar assembly in under 124 seconds and multi-connector insertion within 72 seconds.

Nvidia emphasized the need for flexible automation systems to address the uncertainties in the assembly process, given the rapid design cycles and relatively low production volumes. The robots had to be adaptive and robust enough to handle variations in the assembly process. Hon Hai, a major supplier for Nvidia, announced last year its collaboration to build a smart manufacturing plant in Houston, Texas, using Nvidia's model-driven humanoid robots to produce AI servers.

Hon Hai plans to expand the use of these robots to other plants in Texas, Wisconsin, and California. The company, which dominates over 40% of the global AI server market, predicts that its AI rack shipments will more than double by 2026. Hon Hai's aggressive diversification strategy, known as '3 plus 3,' includes expanding into electric vehicles, robots, and digital healthcare, leveraging AI, semiconductor, and communications technologies.