NVIDIA has introduced NVIDIA® Spectrum-XGS Ethernet, a next-generation networking technology designed to unify distributed data centers into massive, giga-scale AI super-factories.
As global demand for AI continues to surge, traditional data centers are hitting power and capacity limits within single facilities. To achieve meaningful growth, enterprises must scale beyond the walls of individual buildings. However, existing off-the-shelf Ethernet networking infrastructure often struggles with high latency, jitter, and unpredictable performance.
Spectrum-XGS Ethernet addresses these challenges by extending the NVIDIA Spectrum-X™ Ethernet platform with “scale-across” capabilities. This innovation establishes a third foundation of AI computing complementing scale-up and scale-out by interconnecting multiple, geographically distributed data centers into a single high-performance AI ecosystem.
“The AI industrial revolution is here, and giant-scale AI factories are the essential infrastructure,” said Jensen Huang, founder and CEO of NVIDIA. “With NVIDIA Spectrum-XGS Ethernet, we add scale-across to scale-up and scale-out capabilities to link data centers across cities, nations and continents into vast, giga-scale AI super-factories.”
Integrated seamlessly into the Spectrum-X platform, Spectrum-XGS Ethernet leverages adaptive algorithms to optimize performance across distances between facilities. Features such as auto-adjusted distance congestion control, precision latency management, and end-to-end telemetry nearly double the efficiency of the NVIDIA Collective Communications Library. This enables faster multi-GPU and multi-node communication, ensuring predictable performance for AI workloads across globally distributed clusters.
Also Read: EdgeConneX & Lambda to Build AI Factory in Chicago
Among the early adopters of this breakthrough technology is CoreWeave, which will use Spectrum-XGS Ethernet to connect its data centers.
“CoreWeave’s mission is to deliver the most powerful AI infrastructure to innovators everywhere,” said Peter Salanki, cofounder and chief technology officer of CoreWeave. “With NVIDIA Spectrum-XGS, we can connect our data centers into a single, unified supercomputer, giving our customers access to giga-scale AI that will accelerate breakthroughs across every industry.”
The Spectrum-X Ethernet networking platform delivers unmatched scalability and bandwidth density compared to conventional Ethernet, making it ideal for hyperscale AI factories and multi-tenant environments. It integrates NVIDIA Spectrum-X switches and NVIDIA ConnectX®-8 SuperNICs, enabling enterprises to achieve ultralow latency, seamless scalability, and transformative performance for building next-generation AI infrastructure.
This announcement builds on NVIDIA’s continued momentum in networking innovation, following the launch of NVIDIA Spectrum-X and NVIDIA Quantum-X silicon photonics networking switches. Together, these solutions are enabling AI factories to connect millions of GPUs across multiple sites while simultaneously reducing energy consumption and operational costs.
