Hollow-core vs. solid-core fiber

As the name suggests, hollow-core fiber (HCF) contains an inner core filled with low-pressure air, surrounded by solid glass. A simple hollow glass tube, however, would not be sufficient to confine light within the air core. To achieve confinement, the inner wall of the hollow glass tube (typically some tens of microns in diameter) is lined with thin glass capillaries, usually four to six nested hollow tubes with around one micrometer thickness. These capillaries reflect light toward the center, and the resulting multiple reflections interfere constructively to prevent light from leaking into the surrounding glass wall of the HCF. 

As a result, HCF relies on interference effects, similar to the colorful patterns seen on thin films of oil on water. In contrast, solid-core fibers (SCFs) guide light using total internal reflection generated by a refractive index difference between the core and cladding.

While this new fiber structure unlocks enormous performance benefits, HCF has long been held back by one challenge: fiber attenuation. For nearly two decades, researchers have worked to refine HCF designs and fabrication techniques, but losses remained too high for real-world deployment. That’s now starting to change. The past few years have seen a rapid reduction in attenuation reported in the literature, bringing HCFs into the spotlight for near-term practical applications.

It’s important to note that all conventional fiber types classified by their physical properties – such as number of modes (single-mode or multi-mode), chromatic dispersion categories (DSF, NZDSF) or effective area – have a solid glass core. Throughout this post, these existing fiber categories will be referred to collectively as solid-core fibers (SCFs) to clearly distinguish them from the emerging class of hollow-core fibers.

What improvements does HCF offer?

• Latency

From the viewpoint of near-term applications, arguably the most important advantage of HCF is its low latency, around 31% lower than SCF, which is determined by the difference in refractive index between the fiber cores where most of the light is confined.

This property makes HCF very attractive for a range of latency-sensitive applications, including data center interconnects, mobile fronthaul and FTTH, to name a few where the impact is greatest.

• Nonlinear effects

Transmission of optical channels through solid-core fiber such as traditional single-mode fiber is limited by the dependency of the refractive index on optical power, which introduces fiber-induced nonlinear effects that degrade channel performance. This imposes an upper limit on the optical power which can be launched into the fiber, thus limiting the optical signal-to-noise ratio (OSNR) and, in turn, the transmission distance or transmission capacity through the fiber.

HCF has negligible nonlinear effects, allowing much higher launch power to offset high loss of long fiber spans and to increase OSNR. Experiments reported at recent conferences have used the highest available amplifier output power, launching up to 37dBm total optical power into HCF, well above the power limit for even the more tolerant high-effective area varieties of SCF. 

This opens promising prospects for higher capacity per fiber or longer transmission distances, provided that fiber attenuation is low enough.

What once held hollow-core fiber back is now rapidly changing.

• Attenuation

Until recently, high attenuation was a major drawback of HCFs and the main obstacle to deployment. A turning point was the first report by Microsoft and the University of Southampton at OFC 2024 of HCF with attenuation below the theoretical limit for SCF. This was followed by new low-attenuation records which currently stand around 0.05dB/km. Obviously, these are research results obtained on relatively short lengths of fiber and commercial HCF manufactured in large volume will have higher attenuation. The attenuation of HCF planned for deployment has not been publicly disclosed as work is ongoing within each manufacturing company to determine a value that is both feasible in production and adequate for the initial network application. One useful data point was offered by YOFC at ECOC 2025, showing a distribution of attenuation over fiber segments totaling 733km where the average was 0.147dB/km and the highest value around 0.3dB/km.

Narrow gas absorption peaks, particularly those induced by CO2 impact transmission distance in the L-band by attenuating some frequencies within the channel bandwidth. This limitation may be resolved by the time the market is ready to deploy HCF in core networks, which require wider operational bandwidth than available in the C-band.

• Chromatic dispersion

Chromatic dispersion (CD) can be very low (e.g., 3 to 5ps/nm/km), removing the need for optical or digital compensation, while avoiding the nonlinear effects that occur in SCF with similarly low dispersion. Reducing the requirement for digital CD compensation in coherent transceivers also serves to reduce DSP complexity and to limit the power consumption, both highly desirable in the main applications of HCF.

• Transmission bandwidth

The bandwidth for low attenuation in HCF is a design-dependent parameter and therefore can be adjusted. Current telecom-oriented designs target low attenuation in the C-band or across C-, S- and L-bands for compatibility with existing WDM equipment. Future developments could shift to longer wavelengths, where wider usable bandwidths could be available, assuming that transceivers and amplifiers are also developed for these spectral regions.

Looking forward to OFC 2026

The strong interest in HCF from research labs around the world, combined with the substantial investments made by industry to advance development and deployment of HCF, will continue to drive rapid progress. Several of today’s technical limitations may look very different even within the next year.

At OFC 2026, HCF will undoubtedly be one of the most closely followed topics. We anticipate new reports demonstrating continued improvements in transmission performance, broader examples of real-world applications, as well as further progress in the supporting ecosystem – particularly in cabling, splicing, field installation practices and monitoring tools essential for large-scale deployment.

Key questions likely to be addressed include:

• What attenuation levels are now achievable at scale? 
• What span lengths, link distances and total capacities have been demonstrated using commercial-grade amplifiers and transceivers? 
• How close are manufacturers to offering standardized cabling and interconnect solutions? 
• What progress is being made in early field deployments, and what lessons are emerging from these trials?

I look forward to hearing your views on HCF. If you’re attending the OFC conference, stop by the Adtran booth (1001) to continue the discussion in person. And stay tuned for a follow-up after OFC, where I’ll highlight the most important new results and insights presented at the conference.

If you’re heading to OFC and would like to learn more about the topic, please book a meeting through our OFC 2026 page.