China's hollow-core fiber trial hits 51 Tbps over 200 km — no repeaters needed
China Telecom, fiber maker YOFC, and Dekoli have set a new world record: 51.3 Tbps transmitted across 206.5 km of hollow-core fiber (HCF) with no signal repeaters in between. That's enough bandwidth to move roughly 6,400 4K video streams simultaneously — over a single cable, across a distance comparable to London to Newcastle. The trial used the world's longest commercial cross-border HCF link, marking a clear step from lab experiment to real-world deployment.
Air instead of glass
Traditional fiber optic cables carry light through a solid glass core. Glass has a refractive index that slows photons to about 70% of their speed in a vacuum — a physical ceiling you can't engineer around. Hollow-core fiber replaces that glass with air. Light travels faster, and the signal degrades less over distance. The result, per the YOFC 51.3 Tbps field trial, is roughly 31% lower latency compared to conventional fiber — a gap that matters enormously in financial trading, cloud computing, and AI workloads where every millisecond has a price tag.
The 206.5 km span was completed using only standard EDFA amplifiers at each endpoint. No intermediate stations. For network operators, that means fewer active components to maintain, lower running costs, and fewer points of failure across long backbone routes.
The race for long-haul HCF
China isn't alone in this push. Microsoft has already deployed 1,280 km of HCF in live Azure infrastructure with zero reported field failures, and is targeting 15,000 km by late 2026, per its Corning and Heraeus partnerships. Corning is scaling up US production; Heraeus covers European manufacturing. That positions Microsoft's data center interconnects — and by extension the UK and US cloud hubs that rely on Azure — to benefit from lower-latency backbone routes within the next two years.
The China Telecom trial raises the competitive stakes. Achieving 1.2 Tbps per wavelength across 43 channels, unrepeatered, is the kind of result that accelerates procurement decisions across the industry. High-frequency trading routes between exchanges and colocation centers — the Shenzhen-to-Hong Kong model — are the most immediate commercial candidates.
What's still missing
HCF isn't a drop-in upgrade. The ITU-T standards body is still reviewing technical specifications for the technology, meaning carriers can't yet buy off a standardized parts list. Splicing HCF requires new tools and trained technicians. Manufacturing capacity is ramping, but lead times remain long. For US and UK network operators, the technology is now clearly proven at scale — the bottleneck is the supply chain, not the physics.
