A key to NVIDIA’s ability to translate AI chip designs into large-scale hardware products is its collaboration with global manufacturers, investment in new supply chains, and stronger production scheduling. Deb Shoquist explained that the team reduced the time from chips coming off the line at TSMC to final system assembly at Foxconn and Wistron from five weeks to two weeks. It also optimized automated testing, production-line design, and cross-team engineering collaboration around the NVIDIA GB200 compute board.
Source:NVIDIA 官方博客 · Read Original
Leave a Reply