Nvidia to Launch Next-Gen AI Chip Platform in 2026 with Enhanced Networking and Inference Capabilities

Nvidia to Launch Next-Gen AI Chip Platform in 2026 with Enhanced Networking and Inference Capabilities

NVIDIA has announced the widespread adoption of its Spectrum-X platform, designed to enhance Ethernet networking for generative AI. This new technology promises to boost generative AI network performance by 1.6 times compared to traditional Ethernet fabrics. Several cloud service providers, including CoreWeave, GMO Internet Group, Lambda, Scaleway, STPX Global, and Yotta, have already embraced the platform. Additionally, partners such as ASRock Rack, ASUS, GIGABYTE, Ingrasys, Inventec, Pegatron, QCT, Wistron, Wiwynn, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro have announced Spectrum-based products.

In related developments, NVIDIA has launched the general availability of NVIDIA AI Enterprise-IGX with Holoscan, aimed at enabling real-time AI computing at the industrial edge. This platform offers enterprises enhanced performance, security, and support for deploying AI applications at scale. Companies like Medtronic, SETI Institute, and various manufacturing leaders are utilizing this platform to develop innovative solutions. Furthermore, a range of companies, including Advantech, ADLINK, Aetina, and more, are building NVIDIA-Certified IGX systems.

CEO Jensen Huang also announced that NVIDIA's next-generation AI chip platform, Rubin, will be launched in 2026. The Rubin family of chips will include new GPUs, CPUs, and networking chips, and is set to include a new CPU named Versa, along with graphics chips that feature high-bandwidth memory from SK Hynix, Micron, and Samsung. This marks an increase in the frequency of new product releases, moving to an annual schedule from the previous biannual one.

Summary

Other news in technology