Tying multiple data centers together into an integrated whole is a hot topic in data center circles. The goal is to achieve performance beyond what a single data center or AI factory can deliver. But with power demands rising, integrating data centers across multiple location where there’s available power makes sense.
But combining data centers into a massive combination while overcoming the networking limitations of off-the-shelf Ethernet — problems such as jitter, high latency and erratic performance — is the challenge.
Ahead of next week’s Hot Chips 2025 conference in Palo Alto, NVIDIA today announced Spectrum-XGS Ethernet, which the company calls a scale-across technology for combining multiple, distributed data centers into “unified, giga-scale AI super-factories.”
NVIDIA said the Spectrum-X Ethernet networking platform provides 1.6x greater bandwidth density than off-the-shelf Ethernet for multi-tenant, hyperscale AI factories. It comprises NVIDIA Spectrum-X switches and NVIDIA ConnectX-8 SuperNICs, designed to deliver scalability and low latency.
“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.”
The company said Spectrum-XGS Ethernet offers auto-adjusted distance congestion control, latency management and end-to-end telemetry. The offering nearly doubles the performance of the NVIDIA Collective Communications Library, according to NVIDIA, accelerating multi-GPU and multi-node communication to deliver predictable performance across geographically distributed AI clusters. As a result, multiple data centers can operate as a single AI super-factory, fully optimized for long-distance connectivity.
The new offering also builds on NVIDIA’s partnership with AI hyperscaler CoreWeave, which NVIDIA said will be among the first to connect its data centers with Spectrum-XGS Ethernet.
“CoreWeave’s mission is to deliver the most powerful AI infrastructure to innovators everywhere,” said Peter Salanki, co-founder and CTO 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.”
