WhiteFiber, Inc. (NASDAQ: WYFI) has made its new cross-data-center networking architecture, Continuum, commercially available, describing it as the first distributed GPU supercluster system of its kind on the market. The technology links geographically separated data centers into a single logical GPU cluster, aiming to remove bandwidth and distance as constraints on how AI infrastructure is deployed.
Continuum is built on two HITRUST-certified QTS facilities located 83 kilometers apart, connected over 12 Zayo dark fiber strands into one logical system. The architecture is powered by DriveNets AI Fabric for networking and WEKA NeuralMesh for data storage and memory, and delivers 136 Tbps of bandwidth with 0.9 milliseconds of guaranteed round-trip latency. WhiteFiber has submitted patent applications covering the underlying implementation.
The company says the approach lets fragmented telecom and metro sites, which often sit on unused power and fiber capacity, contribute to one pooled GPU supercluster rather than operating as isolated capacity. It's also designed to keep sensitive workloads within their originating jurisdiction for data sovereignty purposes, while allowing training runs to continue if one site goes down.
"WhiteFiber Continuum is the infrastructure answer to a problem the industry has lived with for years: geographic distance as a ceiling on what a cluster can do," said WhiteFiber CEO Sam Tabar in the announcement. DriveNets and WEKA both contributed executive quotes describing their roles in the joint architecture. WhiteFiber Continuum became commercially available on September 23, 2026.