Fiber for Breakfast Week 27: Optical Fiber and Cable Market Outlook: The Role of Hollow-Core and Multicore Fiber

Fiber for Breakfast Week 27: Optical Fiber and Cable Market Outlook: The Role of Hollow-Core and Multicore Fiber 

The fiber industry has been through several waves of growth. Each one has been tied to a major technology shift, from mobile networks to FTTH and now to AI-driven data centers. This week’s Fiber for Breakfast focused on what comes next as those demands begin to push optical networks in new directions.  

Gary sat down with Ahmed Ali, Senior Analyst and Lead of CRU’s Optical Fibre and Cable Services, to discuss the latest market outlook for optical fiber and cable, including the growing role of hollow-core and multicore fiber. The message was clear: conventional single-mode fiber remains the foundation of global connectivity, but some emerging AI applications are driving the need for ultra-low latency and high fiber count cables.  AI Data Centers provide ideal applications for significantly reducing the size on fiber count cable using multicore fiber.  Low latency is also critical in AI Data Centers, providing the perfect application of hollow-core fiber, where low latency of fiber is taken to the next level of ultra-low latency. 

Ali described a market entering a new growth phase, riven by continued FTTH expansion and an even faster rise in data center demand. Telecom applications still represent the majority of global fiber demand, but data centers are growing much faster, fueled by AI workloads, AI expansion, and data center interconnect. CRU expects global optical cable demand to grow this year, with North America leading that growth because of strong data center investment, continued FTTH deployment, and future BEAD-related demand. That is where the story starts to change.  

AI is not only creating demand for more fiber. It is also creating demand for different kinds of performance. Distributed AI architectures require high-speed connections between GPU clusters, data centers, and server environments. That puts new pressure on reducing latency, while improving signal quality, capacity, density, and the physical space constraints available for increasingly large fiber counts.  

Hollow-core and multicore fiber are being developed to address those specific challenges. Hollow-core fiber replaces the solid glass core with air, allowing light to propogate faster and thereby reducing latency. That makes it especially interesting for applications where ultra-low latency matters, including data center interconnect, financial networks, certain enterprise applications and potentially quantum networking. Multicore fiber addresses a different problem by placing multiple cores inside a single fiber strand, increasing capacity density while reducing space and weight in high-density environments.  

But Ali was careful not to overstate where the market is today. “I don’t expect hollow-core or multicore to take over,” he said. “But they will have their space in terms of the challenges they solve.” 

That distinction matters. These technologies are not replacing single-mode fiber in broadband access networks, and they are not likely to become a part of a typical fiber-to-the-home build anytime soon. Instead, they are emerging as premium solutions for specific applications where economies make sense.  

The barriers are still significant. Hollow-core fiber remains difficult to manufacture, with low yields and is expensive to manufacture. Multicore fiber may be somewhat easier to scale, but it still requires progress around manufacturing, splicing, connectors, testing, transceivers, and standards. The broader ecosystem has to mature before either technology can move from trials to more widely available commercial products.  

Still, the market is beginning to organize around them. Ali pointed to growing activity among major fiber manufacturers, hyperscalers, and technology partners, with multicore beginning to see more alignment around design and hollow-core attracting attention for latency-sensitive use cases. CRU expects early volumes to remain limited through the end of the decade, with broader acceleration possible after 2030 if AI and data center demand continue on their current path.  

Gary connected the discussion to the broader “super investment cycle” now shaping the fiber industry. FTTH investment is still underway, but layered on top of that is AI infrastructure, data center growth, quantum networking, and the need for lower-latency, higher-capacity optical. Networks. 

The takeaway is not that fiber providers need to rethink everything they are building today. It is that the next phase of growth may require a broader toolbox. Single-mode fiber remains the workhorse, but hollow-core and multicore fiber point to where the market is heading as AI pushes networks into more specialized, performance-driven environments. 

Click here to watch the full interview.  

Click here to view the slides.