WhiteFiber Achieves 111.2 Tbps Bandwidth in Cross-Data-Center AI Infrastructure Test
Event summary
- WhiteFiber's cross-data-center networking solution achieved 111.2 Tbps bandwidth across 83km with sub-millisecond latency in initial R&D testing.
- The architecture links two geographically separated data centers into a single logical GPU supercluster, enabling AI workloads beyond the capabilities of any single facility.
- Commercial launch is targeted for Q3 2026, with full-fiber spectrum testing planned before then.
- WhiteFiber has submitted patent applications for the underlying implementation.
The big picture
WhiteFiber's achievement represents a significant advancement in AI infrastructure, potentially reshaping how enterprises deploy and scale AI workloads. The ability to link geographically separated data centers into a single logical supercluster addresses key constraints around scale, resilience, and compliance. This development comes at a time when demand for high-performance computing solutions is surging, driven by the growth of generative AI and other advanced applications.
What we're watching
- Technical Execution
- Whether WhiteFiber can successfully complete full-fiber spectrum testing and meet the Q3 2026 commercial launch timeline.
- Market Adoption
- The pace at which enterprises will adopt this distributed GPU supercluster architecture for AI workloads.
- Competitive Positioning
- How this innovation will position WhiteFiber against competitors in the AI infrastructure space.
Related topics