New Fiber Tech for AI Infrastructure Unveiled
At MWC Barcelona, Yangtze Optical Fibre and Cable (YOFC) is set to unveil a new Hollow-Core Fibre (HCF) solution. The company claims the next-gen optical tech will offer ultra-low latency communication, aiming to strengthen the global infrastructure required to support the massive data demands of AI.
Hollow-core fibre guides light through an air-filled channel, unlike traditional solid-core fibers that use glass. This fundamental design difference allows light to travel nearly 50% faster, approaching the speed of light in a vacuum, which can reduce latency by approximately 30-31%. For AI clusters distributed across multiple data centers, this reduction in signal delay significantly expands the viable distance between facilities. The technology nearly eliminates the Kerr optical nonlinearity found in solid-core fibers, enabling higher power handling and improving signal fidelity. It also boasts backscattering that is 10,000 times lower than traditional fibers, reducing signal attenuation. At the OFC Conference in March 2025, YOFC announced it had achieved a record-low attenuation of 0.05dB/km and could manufacture single fibers over 20 kilometers in length. Manufacturing hollow-core fiber is significantly more complex and costly than producing traditional fiber. The process requires precise control to maintain the uniform, thin-walled tube structure and prevent nanometer-scale imperfections that can scatter light. Splicing and connecting these fibers also demand specialized techniques to align their complex microstructures. In June 2024, YOFC, in partnership with China Mobile, launched the world's first 800G hollow-core fibre transmission test network in Guangdong. This real-world deployment was designed to evaluate the fiber's performance under operational conditions, including challenges like pipe flooding and outdoor splicing, and achieved a throughput of 128Tb/s over the test link. This initiative is part of YOFC's broader "AI-2030" strategy, positioning the company to provide core infrastructure for the massive data demands of AI.