WhiteFibre launches cross-data-centre AI supercluster

WhiteFibre has launched WhiteFibre Continuum, a cross-data-centre networking system that links separate facilities into a single GPU supercluster. The product is now commercially available to enterprise customers.

The system connects two HITRUST-certified QTS data centres 83 kilometres apart and presents them as one logical cluster. It runs over 12 strands of Zayo dark fibre and uses DriveNets AI Fabric with WEKA NeuralMesh storage infrastructure.

The launch addresses a broader challenge for AI infrastructure builders as demand for compute collides with limits on power, land, and permitting for new single-site campuses. WhiteFibre is positioning the product as an alternative to expanding capacity only through large new developments.

According to the company, Continuum delivers 136 Tbps of bandwidth with 0.9 milliseconds of guaranteed round-trip latency across the two sites. WhiteFibre has also filed patent applications covering the underlying implementation.

The product builds on testing WhiteFibre disclosed earlier, which validated 111.2 Tbps of bandwidth and 0.9 millisecond round-trip latency over the same 83-kilometre span. According to the company, that result came within 8% of the physical limit for light travelling through fibre over that distance.

Distributed model

WhiteFibre says the architecture is designed to reduce reliance on a single location for large-scale AI workloads. In practice, compute resources can be pooled across sites, and workloads can continue running if one facility goes offline.

The company is also targeting customers that face rules on where data can be stored and processed. Sensitive workloads can remain within their original jurisdiction while failover and shared compute operate across multiple locations, according to WhiteFibre.

That approach could appeal to businesses in finance, healthcare, and government, as well as operators in telecommunications, edge computing, and sovereign AI projects. For many of those organisations, resilience and data-location rules are central to infrastructure design.

Sam Tabar, Chief Executive Officer of WhiteFibre, framed the release as a response to a long-standing technical constraint. "WhiteFibre 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," Tabar said.

"Today we are making it commercially available to enterprises that need AI compute to perform at scale, hold up under compliance requirements, and not break when a single site has a problem. This is the foundation for a new class of AI infrastructure, and we are ready to build it with our customers," he added.

Partner roles

WhiteFibre built the offering with DriveNets and WEKA, which remain part of the commercial product. DriveNets supplies the network fabric between the sites, while WEKA provides the data storage and memory layer across the combined cluster.

Yossi Kikozashvili, Vice President of Product and GTM - AI Infra at DriveNets, said the networking design was central to making the setup work across distance. "WhiteFibre's Continuum shows what becomes possible when ambitious engineering is paired with the right networking architecture," Kikozashvili said.

"DriveNets' Ethernet-based AI fabric delivers the highest performance even in the most demanding, high-bandwidth, low-latency environments, ensuring scale-across superclusters move data efficiently, maximize GPU utilization, and optimize power efficiency," he added.

WEKA focused on the challenge of maintaining shared access to data across physically separate facilities. "The industry has treated distance as a hard ceiling on AI infrastructure, and WhiteFibre Continuum breaks it. But a supercluster is only as unified as its data. WEKA's NeuralMesh gives Continuum a single, high-performance storage and memory foundation across sites, so GPUs that sit many kilometers apart operate as if the data were local. This is what turns two data centers into one engine, and what the next generation of AI innovation will be built on," said Liran Zvibel, Co-Founder and Chief Executive Officer of WEKA.

The launch comes as infrastructure providers look for ways to use existing assets more efficiently rather than wait for new large-scale sites to clear planning and utility hurdles. WhiteFibre says metro and telecom facilities with available power and fibre could be brought into a shared cluster model instead of operating as isolated capacity.

That argument is likely to resonate in markets where grid access and construction timelines are becoming decisive factors in AI deployment. By linking existing facilities into one logical system, providers may be able to expand available compute without relying entirely on a single massive campus.

For WhiteFibre, the commercial release marks a shift from research results to a product aimed at enterprise deployment, built on the premise that distance between data centres no longer has to define the limits of an AI cluster.