NVIDIA has announced significant adoption of its Spectrum-X Ethernet networking platform by leading AI cloud service providers including CoreWeave, GMO Internet Group, Lambda, Scaleway, STPX Global, and Yotta. These companies are integrating Spectrum-X to enhance their AI infrastructures’ networking performance. 

Additionally, several NVIDIA partners such as ASRock Rack, ASUS, GIGABYTE, Ingrasys, Inventec, Pegatron, QCT, Wistron, and Wiwynn are incorporating Spectrum-X into their products. They join established tech giants Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro in adopting this platform.

Gilad Shainer, NVIDIA’s Senior Vice President of Networking, emphasized the importance of networking innovation for businesses leveraging generative AI. Spectrum-X, the world’s first Ethernet fabric designed for AI, would enhance generative AI network performance by 1.6 times compared to traditional Ethernet fabrics.

AI Ethernet networking

During his COMPUTEX keynote in Taiwan this week, NVIDIA founder and CEO Jensen Huang announced an accelerated release schedule for Spectrum-X products, with new versions launching annually. These updates will include increased bandwidth, more ports, and advanced software features to maintain AI Ethernet networking leadership.

The Spectrum-X platform, featuring the Spectrum SN5600 Ethernet switch and BlueField-3 SuperNIC, offers adaptive routing, congestion control, and high-performance Ethernet networking suitable for AI workloads. It would ensure reliable performance for thousands of simultaneous AI tasks. Paired with NVIDIA BlueField-3 DPUs, Spectrum-X would facilitate advanced cloud multi-tenancy, GPU compute flexibility, and robust security, enabling AI cloud service providers to boost development and deployment while enhancing ROI.