Introduction

For decades, optical fibers have relied on a solid glass core to guide light and have formed the backbone of global telecommunications. However, glass imposes a fundamental physical limitation because light travels through it approximately 30 percent slower than through air. In standard silica fiber, the group velocity of light is about 2×10⁸ meters per second, approximately 67% of the speed of light in vacuum, which results in a latency of around 5 microseconds per kilometer. This constraint has long been accepted as a trade-off for the reliability and manufacturability of solid-core fibers. But this is where hollow-core fiber begins to change the landscape. By replacing the solid core with an air-filled channel, hollow-core fibers (HCFs) allow light to propagate at nearly its vacuum speed, reaching approximately 3×10⁸ meters per second. This reduces latency to around 3.3 to 3.5 microseconds per kilometer, offering a 30 to 50 percent speed increase over traditional fibers. Beyond speed, hollow-core fibers drastically reduce glass interaction by factors ranging from 10² to 10⁵, which virtually eliminates nonlinear effects such as Kerr and Brillouin or Raman scattering. They also support a broader low-loss optical spectrum that can span from the visible range to approximately 2100 nm and may even achieve lower attenuation than conventional single-mode fibers by avoiding the Rayleigh scattering limit of silica. This shift marks the emergence of hollow-core fiber as a transformative technology and invites a deeper exploration of its design principles, performance characteristics, and deployment challenges.

This article begins with a review of the historical development of hollow-core fibers and the key milestones that have led to today’s advanced designs. It then outlines the theoretical principles behind HCF technology and describes the main types of hollow-core fibers along with their respective advantages and limitations. A comparison between solid-core silica fibers and hollow-core fibers is presented, focusing on telecom-relevant metrics. The article concludes with a summary of current challenges and outlook.

Hollow-core fiber innovations and milestones

The technological timeline of hollow-core fibers is illustrated in Figure 1. Prior to the year 2000, researchers explored silica-air capillaries for applications such as gas sensing, interferometric measurements, and atom guidance. However, these early designs suffered from high attenuation in simple glass tubes, which limited their practical use. The first major leap arrived in 1999 when Cregan et al. at the University of Bath demonstrated a hollow-core photonic bandgap fiber (HC-PBGF), proving that a microstructured cladding could confine light in an air core [1]. From that point, HCF development diverged into two routes: photonic bandgap guidance using periodic lattices and anti-resonant or inhibited-coupling designs based on thin-walled tubes. The telecom-relevant story begins in 2002 with a 7-cell HC-PBGF reporting 13 dB/km [2], followed by the first Kagome HCF [3] designs (then < 2000 dB/km, broad-band but leaky). Steady structural refinements drove losses down: a 13-cell PBGF reached 1.2 dB/km (2005) [4]; introducing negative curvature around 2010 cut Kagome-type losses to < 250 dB/km [5]. In parallel, simpler tubular ARF geometries emerged—single-ring “revolver” fibers were shown by 2011 (< 1000 dB/km) [6], improving to 7.2 dB/km (2017) [7]. A co-joined-tube ARF hit 2 dB/km (2018) [8], and the breakthrough came with nested anti-resonant nodeless fibers (NANF), which suppressed leakage with inner “nesting” rings and delivered 0.28 dB/km (2020) [9]. Together, these milestones mark the transition from narrow-band PBG guidance to broadly useful, low-loss anti-resonant designs suited for telecom.

Today, hollow-core fiber is transitioning from laboratory research to real-world deployment. Microsoft’s Azure team has already demonstrated its potential in AI and cloud infrastructures, reporting significant reductions in latency and measurable performance gains across data-center fabrics. Major telecom operators have also conducted field trials, validating the technology under operational conditions. At the same time, specialized manufacturers are now delivering cable-grade NANF designs, engineered for splicing, installation, and long-haul reliability. These steps collectively mark the shift of HCF from an experimental platform to an emerging commercial technology positioned to reshape next-generation networks.

Principles of hollow-core fiber

In conventional optical fibers, light confinement is achieved through the principle of total internal reflection (TIR). Here, the fiber core has a higher refractive index than the cladding, ensuring that light entering at shallow incidence angles is reflected back into the core and guided along the fiber. For further details, please refer to our recent article [11]. In hollow-core fibers, however, the situation is reversed: the core is filled with air (n≈1) and the cladding is typically silica glass (n≈1.45), so the condition for TIR cannot be satisfied. Instead, HCFs employ more advanced mechanisms to confine light within the hollow core. Three major approaches have been successfully implemented: photonic bandgap (PBG), Bragg, and anti-resonant (or inhibited coupling), as illustrated in Figure 2.

Photonic bandgap fiber

The photonic bandgap effect arises from a periodic dielectric structure that prohibits the propagation of certain frequency ranges, much like electronic bandgaps in semiconductors restrict electron energies. In HCFs, the cladding is engineered as a two-dimensional photonic crystal, typically consisting of a lattice of air holes in silica, while the hollow core acts as a defect in the lattice. Light whose wavelength falls within the photonic bandgap is forbidden from propagating in the cladding and is therefore confined to the hollow defect. The physics of this phenomenon is governed by Maxwell’s equations in periodic media, leading to the eigenvalue problem

where μ(r) is the magnetic permeability distribution, ε(r) is the spatially varying dielectric function, E(r) is the electric field, ω is the angular frequency, and c is the speed of light in vacuum. This eigenvalue relation implies that only certain frequencies (or equivalently, wavelengths) are allowed within the crystal structure, while others are reflected into the hollow core. Photonic bandgap fibers (PBGFs) are practical realizations of this principle. They typically feature a hexagonal lattice of air holes surrounding a central hollow core. These fibers can achieve low attenuation and single-mode operation within the bandgap, but their guidance bandwidth is relatively narrow (often <50 nm), and performance degrades sharply outside this range. Fabrication requires precise control over hole size, spacing, and symmetry, making them sensitive to structural imperfections.

Bragg fiber

Bragg fibers are another implementation of PBG, but instead of a 2D lattice, they use a one-dimensional (1D) radial structure. The cladding consists of concentric layers of alternating high- and low-refractive-index materials, forming a cylindrical Bragg mirror. When the layer thicknesses satisfy the quarter-wave condition

where ni and di are the refractive index and thickness of each layer, constructive interference reflects specific wavelengths (λ) back into the core. This structure confines light within a narrow spectral band. Bragg fibers offer strong mode confinement and can be single-mode even with large core diameters. However, they suffer from limited bandwidth and high fabrication complexity. Their performance is highly sensitive to layer uniformity and refractive index contrast. While effective for specific wavelengths, they are less suitable for broadband or telecom applications due to narrow guidance windows and higher attenuation.

Anti-resonant fiber

An alternative approach is anti-resonant reflection, which relies on silica thin-walled capillaries surrounding the hollow core. When the wall thickness d satisfies the antiresonance condition

where n_glass is the refractive index of the silica, λ is the wavelength of light, and m is an integer, destructive interference suppresses transmission into the cladding while constructive interference enhances reflection. This effect is analogous to a Fabry–Pérot etalon, but exploited in reverse: instead of maximizing transmission, it minimizes it, thereby confining light to the hollow core. The resulting broadband reflection enables anti-resonant fibers (ARFs) to guide light across spectral ranges spanning hundreds of nanometers. ARFs include several subtypes:

• Kagome fibers: use a web-like cladding structure with thin silica struts arranged in a hexagonal symmetry.
• Single-ring ARFs: consist of one ring of capillaries around the core, often with negative curvature to inhibit coupling.
• Nested anti-resonant nodeless fibers (NANFs): feature additional internal capillaries nested within the primary tubes to further suppress leakage and achieve ultra-low loss. NANFs currently lead in performance, achieving attenuation below 0.2 dB/km and bandwidths exceeding 150 nm. They are well-suited for telecom due to their low latency, minimal nonlinearity, and compatibility with high-speed data transmission.

Comparing HCF designs

Photonic bandgap and anti-resonant fibers represent two distinct approaches to hollow-core guidance, each with trade-offs. PBGF initially achieved lower losses but within narrow transmission windows—typically tens of nanometers. Their bandgap structure inherently filters out-of-band light, which is useful for nonlinear optics and spectral shaping. However, their small core size (10–15 µm) can lead to higher nonlinearity at elevated powers.

ARFs, including NANFs, offer much broader bandwidth—often hundreds of nanometers—and have recently surpassed PBGFs in loss performance. Their larger cores support higher power transmission with lower nonlinearity, making them ideal for ultrabroadband and high-capacity telecom links. While photonic bandgap fibers require precise periodic microstructures over long lengths, anti-resonant fibers demand accurate control of tube geometry but involve fewer structural elements overall. As of 2023, industry consensus favors AR/NANF fibers for telecom due to their superior combination of low loss, wide bandwidth, and scalability. PBG fibers remain valuable in specialty applications but are no longer the leading candidate for mainstream deployment. For a side-by-side comparison of the three fibers, see Table 1.

Telecom applications and outlook

Ultra-low latency with hollow-core fiber

In high-performance networks, latency isn’t just a metric—it’s a competitive advantage. Whether you're operating a financial trading platform, managing real-time AI workloads, or synchronizing data centers across regions, every microsecond counts. This is where ARFs redefine what’s possible. Unlike conventional silica fibers, where light travels at ~67% the speed of light in vacuum (due to the refractive index of glass, n ≈ 1.45), ARFs guide light through air (n ≈ 1.0003), allowing it to propagate at nearly 99.7% of the speed of light. This results in a latency reduction of approximately 30% over the same distance.

For example, over a 40 km link, ARF can deliver data nearly 48 microseconds faster than standard fiber. In latency-critical environments like high-frequency trading, this time difference can translate into millions of dollars in advantage. In cloud infrastructure, it means faster synchronization, lower jitter, and better responsiveness for distributed computing.

High-capacity transmission

With losses now in the <0.2 dB/km range, HCFs are viable for longer distances. Recent research demonstrated wavelength-division multiplexing (WDM) signals over 1000+ km in NANF fiber with standard amplifiers and no fundamental issues. The greatly reduced nonlinearity means HCFs can potentially carry higher power or more channels before hitting nonlinear Shannon limits. Also, hollow fibers have much lower chromatic dispersion (and slope) in C-band – in fact a DNANF shows ~7× lower dispersion than SMF – which could simplify DSP for coherent systems or allow launching wider spectra without dispersion compensation.

Figure 3 illustrates the comparative loss profiles of DNANF and two benchmark silica fibers: the 2002 low-loss fiber from Nagayama et al., and the 2025 PSCF from Sato et al. The DNANF exhibits a mean attenuation below 0.14 dB/km across a 424 nm bandwidth centered at 1,504 nm, equivalent to 54.3 THz. Within this range, a narrower window of 144 nm around 1,553 nm shows attenuation below 0.1 dB/km, highlighting the fiber’s suitability for dense wavelength-division multiplexing (DWDM) and high-capacity coherent transmission.

This broad low-loss region is particularly significant for telecom systems, as it enables the use of wider optical spectra without the need for frequent amplification or dispersion compensation. Compared to silica fibers, which typically offer low loss over narrower bands, DNANFs provide both spectral flexibility and latency advantages due to their air-guided core.

Specialty environments

HCFs have no or minimal glass in the core, so they are very attractive for situations where standard fibers struggle:

• High-power delivery: For optical power delivery (e.g. for fiber lasers or LIDAR), an HCF can carry kilowatts of power with negligible nonlinearity and large mode area, without worrying about glass damage or Brillouin scattering.
• Sensing and metrology: The air core can be filled with gases for distributed sensing or provide stable propagation for precision interferometry and frequency comb distribution (hollow fibers have less thermal sensitivity and virtually no Fresnel reflections at the core interface).
• Radiation-hard and low thermal noise links: In radiation environments (nuclear facilities, space), hollow cores avoid the radiation-induced attenuation that affects doped silica. Also, HCF latency is very stable with temperature changes (since index of air is barely affected, unlike silica’s index), which can benefit radio antenna synchronization and other timing links.
• Mid-IR and specialty bands: Hollow-core fibers made of silica can guide well beyond silica’s normal transparency window because most light is in air. For instance, HCFs can operate at 2000+ nm (well into the infrared) while standard SMF suffers high loss there. This opens possibilities for using new wavelengths (like 2 µm band) in communications or for delivering CO2 laser light (10 µm) with an all-dielectric fiber (OmniGuide’s original use).

Limitation and challenges

Manufacturing & cost: HCFs are currently more expensive to produce than SMF, due to complex preform fabrication and lower yields. As techniques improve (e.g. automatic stack-and-draw for PBGFs or improved tube stacking for NANFs) and volume increases, costs should drop. The establishment of dedicated hollow-fiber factories (like Lumenisity’s facility in UK) is a step in this direction.

Fiber handling: Hollow fibers, especially with thin membranes, can be more fragile under stress and bends. Cable designs need to protect them from sharp bends and crushing. Initial deployments use robust encapsulation to ensure reliability comparable to standard fiber cables.

Connectorization and splicing: Joining an HCF to standard fiber without big losses or reflections is tricky – the air/glass interface reflects ~4% of light. Current solutions include specialized connectors or splicing with an angled/anti-reflection interface or using intermediate graded-index fibers as mode adapters. As these methods mature, we’ll see seamless integration of HCF segments in networks (e.g. hollow-core pigtails for transceivers, or fusion splicers tailored for HCF).

Standardization: Telecom operators will require standardized fiber types (with specified dimensions, reliability, etc.). Work is ongoing to define standard HCF cable specs and to qualify their long-term environmental stability (e.g. ensuring no moisture ingress that could add attenuation).

Despite challenges, the momentum behind hollow-core fiber development is strong. Recent breakthroughs have essentially validated that HCFs can meet or exceed the performance of solid fibers in critical metrics. The telecom industry now recognizes HCFs not just as lab curiosities but as potential enablers for next-generation networks. For instance, a hollow-core fiber that is 47% faster and lower loss than SMF could reduce the need for amplifiers (lower cost) and cut latency (higher performance) in wide-area networks. Companies like Microsoft, British multinational telecommunications (BT), Comcast, and others publicly sharing trial results underscores this potential.

Conclusion

Hollow-core fibers have evolved from a theoretical idea to a practical technology on the cusp of commercialization in telecom. Bragg fibers and photonic bandgap fibers paved the way, demonstrating the fundamental science of bandgap guidance. Today, anti-resonant hollow-core fibers are taking the torch, shattering loss records and showing that guiding light in air can unlock performance beyond what solid glass fibers allow. As fabrication techniques mature and deployment challenges are resolved, we can expect to see HCFs complementing or even replacing conventional fiber in specialized high-performance routes – bringing us closer to the vision of light traveling through networks virtually as fast as it does in free space, but with the reliability and controllability of fiber optics. The next few years will be exciting as lab results translate into real-world infrastructure, potentially marking a major shift in optical communication technology.

Technologie Optic Inc. recognizes the transformative potential of hollow-core fiber technology and is actively investing in research, prototyping, and strategic partnerships to accelerate its adoption. Our engineering team is committed to advancing HCF integration into next-generation telecom infrastructure, with a focus on ultra-low latency, high-capacity transmission, and scalable deployment. We believe hollow-core fibers will play a central role in shaping the future of optical communication, and we are excited to contribute to that evolution.