Researchers at the University of Southampton, working with Microsoft Azure Fibre engineers, have demonstrated a hollow‑core optical fiber that sets a new laboratory record for low attenuation and promises markedly lower latency and higher spectral capacity than conventional silica single‑mode fiber—while also exposing a realistic list of engineering and deployment challenges that must be solved before the technology becomes ubiquitous.
Hollow‑core fibers (HCFs) replace the glass core of traditional telecom fiber with an air‑filled channel and confine light using a microstructured arrangement of thin glass membranes. The physical payoff is straightforward: light guided mostly through air experiences far less Rayleigh scattering and a group index much closer to vacuum, producing lower per‑kilometre loss and reduced propagation delay. The recent work—described by the University of Southampton and reported by multiple trade outlets—implements a Double Nested Antiresonant Nodeless Fiber (DNANF) geometry that nests ultra‑thin glass capillaries to produce anti‑resonant confinement with very low leakage.
Laboratory measurements on multi‑kilometre spools of DNANF show a minimum measured attenuation of 0.091 dB/km at 1,550 nm, a performance point that undercuts the practical floor long‑associated with the best silica fibers (roughly 0.14–0.16 dB/km) and marks what the authors and industry press describe as the lowest signal loss yet recorded for an optical fiber.
That said, turning a laboratory breakthrough into the global backbone of the internet requires disciplined engineering and ecosystem work: scaling manufacturing yields, certifying splices and connectors, proving long‑length reel performance under field conditions, and aligning photonics vendors to exploit broader spectral windows. The most likely near‑term trajectory is selective, high‑value deployment—data‑center interconnects, trading links and specialized scientific or sensing applications—followed by gradual expansion as production, cost and interoperability metrics mature.
Hyperscalers and carriers planning network roadmaps should adopt pragmatic optimism: invest in pilots where the benefits are compelling, require independent validation and operational acceptance tests, and negotiate supplier commitments that reflect the realities of bringing a new fiber class from lab to long‑haul operations. The DNANF milestone is a major technical achievement—and its commercial significance will be determined by the quality of follow‑through engineering and standards work over the coming years.
Conclusion
DNANF hollow‑core fiber is not a marketing fantasy: it is an experimentally validated advance built on decades of hollow‑core physics and years of focused R&D. The technology promises lower loss, lower latency and more spectral headroom—advantages that can reshape how cloud and network operators design high‑value routes. The central question now is not whether DNANF works in the lab, but whether industry can translate spool‑level success into repeatable manufacturing, robust field operations and an interoperable ecosystem. The next 24–48 months will decide whether this physics milestone becomes a practical foundation for the network‑era demands of AI, real‑time services and hyperscale cloud.
Source: Optics.org Southampton achieves ‘record’ low signal loss with hollow fiber
Hollow‑core fibers (HCFs) replace the glass core of traditional telecom fiber with an air‑filled channel and confine light using a microstructured arrangement of thin glass membranes. The physical payoff is straightforward: light guided mostly through air experiences far less Rayleigh scattering and a group index much closer to vacuum, producing lower per‑kilometre loss and reduced propagation delay. The recent work—described by the University of Southampton and reported by multiple trade outlets—implements a Double Nested Antiresonant Nodeless Fiber (DNANF) geometry that nests ultra‑thin glass capillaries to produce anti‑resonant confinement with very low leakage. Laboratory measurements on multi‑kilometre spools of DNANF show a minimum measured attenuation of 0.091 dB/km at 1,550 nm, a performance point that undercuts the practical floor long‑associated with the best silica fibers (roughly 0.14–0.16 dB/km) and marks what the authors and industry press describe as the lowest signal loss yet recorded for an optical fiber.
What the team actually measured
The headline numbers
- Measured minimum attenuation: 0.091 dB/km at 1,550 nm — reported on spool tests and highlighted in the technical publication and press materials.
- Broad low‑loss spectral window: sustained loss below ~0.2 dB/km across tens of terahertz (reported as roughly a 66 THz low‑loss window in the paper and briefings).
- Speed / group‑index advantage: effective propagation closer to vacuum speed, characterized in press and technical notes as approximately 45–47% faster photon transit compared with conventional silica single‑mode fiber (a range that depends on exact fibre geometry and operating wavelength).
Where these numbers came from
The measurements reported are drawn from laboratory characterization of spool lengths and controlled pilot deployments. The authors present detailed attenuation spectra, group‑index measurements and dispersion curves in their paper, and Microsoft/Azure engineering teams have described pilot installations in press materials and technical briefings. Independent trade outlets and university press pages corroborate the major figures, though some deployment metrics cited in corporate statements should be read as operational claims rather than independent audits.Why DNANF matters: physics and system‑level effects
Lower loss → fewer amplifiers, lower energy
Every decibel saved per kilometre translates to a large cumulative difference across hundreds or thousands of kilometres. A fiber with 0.091 dB/km instead of ~0.14 dB/km reduces total attenuation across long spans significantly, which can reduce the number and power of optical amplifiers required for backbone and submarine links. That directly lowers OpEx (power + cooling) and affects CapEx (fewer repeater stations). The research teams and Microsoft have made precisely this argument in framing the technology’s business case.Lower group index → real latency wins
Because the guided mode sits mostly in air, the effective group index is substantially lower than in silica, and photon transit times per kilometre fall accordingly. The result is materially lower one‑way propagation delay—an attractive property for high‑frequency trading, distributed AI inference and other latency‑sensitive services. The published work and Microsoft briefings quantify this advantage in the mid‑40s percentage range versus standard SMF under the tested conditions.Wider usable spectrum → more capacity (if the ecosystem follows)
The DNANF’s wide low‑loss window opens spectral real estate well beyond today’s crowded C‑band. In theory, that increases the potential aggregated bandwidth per fiber by allowing more WDM channels across a broader band. In practice, realizing that capacity requires amplifiers, transceivers and components that operate efficiently outside today’s standard bands—an ecosystem challenge discussed below.Technical anatomy: what makes DNANF work
Antiresonant reflection and nested capillaries
DNANF relies on ultra‑thin concentric silica tubes around the hollow core that act as anti‑resonant mirrors for the wavelengths of interest. The “double nested” structure places tiny capillaries inside larger capillaries to better suppress leakage pathways and reduce coupling to higher‑order cladding modes. Crucially, the design avoids physical contact “nodes” between nested elements—a nodeless geometry that helps limit scattering and leakage.Loss terms addressed
The team targeted three dominant contributors to HCF loss:- Confinement leakage — reduced by tighter anti‑resonant nesting.
- Surface scattering — reduced by exceptional inner‑wall smoothness and process control.
- Microbend‑induced coupling — mitigated through geometry and mechanical design choices.
Dispersion and nonlinearity
DNANF exhibits markedly lower chromatic dispersion in the measured regimes compared with standard SMF, simplifying DSP requirements in coherent receivers and lowering their power use. The air core also reduces nonlinear effects (e.g., Kerr nonlinearity, Raman scattering), which is beneficial for higher launch powers and very high aggregate capacities.Strengths: Where DNANF truly shines
- Record low attenuation: The 0.091 dB/km figure is a milestone for HCF designs and, if reproducible at scale, a genuine operational advantage versus legacy fibers.
- Latency reduction: Nearly halving propagation delay per kilometre in some framing—a physics‑driven advantage—is a hard, direct benefit for latency‑sensitive services.
- Broad, usable spectrum: A wide low‑loss window could unlock new WDM strategies and increase per‑fiber capacity if partners build the necessary components.
- Lower nonlinear penalties: Systems constrained by nonlinearity at high launch powers can benefit substantially.
Risks, engineering hurdles and open questions
The achievement is important, but the path from spool‑level results to production rollouts is nontrivial. Key challenges include:- Manufacturing scale and yield: Producing thousands of kilometres with the tight geometric tolerances and inner‑wall smoothness required for sub‑0.1 dB/km loss requires new process automation, inline metrology, and yield improvements. Lab reels and pilot spools are not the same as continuous production draws.
- Splicing, connectors and field handling: HCF’s nested microstructure demands different splicing methods, connector designs and cable jacketing. Field splicing loss and repeatability are critical acceptance metrics for carriers. Existing ODFs and connector types may not be drop‑in compatible.
- Bend tolerance and cabling robustness: Mechanical protection, microbend sensitivity and moisture/particulate ingress control in the hollow core are real concerns for real‑world duct, aerial and campus deployments.
- Ecosystem readiness: Wider spectrum exploitation requires amplifiers and transceivers across new bands; without coordinated component availability, the raw fiber capacity advantage remains theoretical.
- Independent, long‑term validation: Many published figures derive from controlled tests and internal pilots. Deployment claims (kilometres installed, target rollouts) are corporate statements that deserve independent route‑level verification and long‑duration environmental testing.
Deployment claims: cautious optimism required
Microsoft and partners have framed DNANF as a strategic way to accelerate Azure performance. Public statements and trade coverage reference pilot installations and operational claims—figures cited include ~1,200 km of installed HCF carrying live traffic in pilot configurations and ambitious internal targets such as 15,000 km planned within Azure. These are significant operational claims, but they appear in corporate briefings and press reporting rather than the peer‑reviewed scientific paper itself; therefore they should be treated as company statements until verified by independent audits or carrier acceptance tests.Use cases most likely to adopt DNANF first
The balance of risk/reward points toward selective early adoption in high‑value, controlled environments:- Data center interconnects (DCI) and hyperscaler backbone links where operators control installation and maintenance.
- Financial trading links and other microsecond‑sensitive point‑to‑point routes.
- High‑power laser delivery and specialized sensing where nonlinearity and damage thresholds matter.
- Hybrid cable strategies where HCF strands run alongside conventional SMF in the same sheath to enable progressive migration.
What network architects and procurement teams should do now
- Treat DNANF/HCF as an emerging strategic option for latency‑sensitive or long‑haul private builds, but avoid wholesale replacement of proven SMF plants until operational maturity is confirmed.
- Run controlled pilots that instrument splice yields, OTDR signatures, bend loss, and amplifier interoperability under production conditions. Demand spool‑length attenuation guarantees and provide acceptance tests that reflect installed realities.
- Require vendors to provide environmental aging data, long‑length reel attenuation distributions, and field splice procedures as part of procurement.
- Negotiate contracts that reflect realistic ramp timelines for transceiver and amplifier ecosystems needed to exploit broader spectral windows.
Strategic implications for hyperscalers, carriers and the wider market
If DNANF’s lab figures are reproducible in production and installed deployments, the market implications are broad:- Hyperscaler edge: An operator that can reliably deploy low‑loss HCF across critical Azure‑to‑Azure routes gains a structural latency and cost advantage for AI training, real‑time inference and premium enterprise services. Microsoft’s vertical integration (acquiring Lumenisity, partnering with large manufacturers) reflects that strategy.
- Carrier economics: Fewer amplifiers, lower power profiles and extended amplifier spacing could shift long‑haul and subsea economics—if manufacturing cost per km reaches parity or is offset by operational savings.
- Competition and standards: Multiple vendors and research groups are working on HCF variants; industry standardization on connectors, splicing, test procedures and mechanical qualifications will be essential to prevent fragmentation and vendor lock‑in.
- Timeline: Near‑term pilots and targeted production reels are realistic in the next 12–36 months; broader commercial adoption across public carrier backbones will likely take multiple years as the ecosystem matures.
Caveats and unverifiable claims (flagged)
Several widely circulated statements should be read with caution:- Deployment totals such as 15,000 km planned and specific installed‑km figures are corporate ambitions or internal operational claims and have not appeared as independently audited route maps or regulator filings in the public domain. Treat them as meaningful indicators of intent rather than confirmed, universally validated installation metrics.
- Spectral multiplication (claims like “5–10× more usable bandwidth”) is an engineering projection based on the fiber’s low‑loss window. Realizing such multiplication requires amplifier and transceiver ecosystems to support additional bands at scale—work that remains to be done.
- The headline latency percentage (45–47%) reflects measured group‑index benefits in controlled tests and depends on exact fiber mode confinement and wavelength; real network latency gains also depend on route geometry, switching and device processing delays, so the fiber’s percentage speedup is a physical upper bound, not an end‑to‑end latency guarantee.
Bottom line: a breakthrough that needs pragmatic follow‑through
The DNANF results published by the University of Southampton team with Microsoft Azure Fibre collaborators mark a genuine step change in hollow‑core fiber performance: record low attenuation (0.091 dB/km at 1,550 nm), broad low‑loss windows and measurable group‑index (latency) advantages. These are physics‑backed improvements with clear value propositions—especially to hyperscalers and latency‑sensitive verticals.That said, turning a laboratory breakthrough into the global backbone of the internet requires disciplined engineering and ecosystem work: scaling manufacturing yields, certifying splices and connectors, proving long‑length reel performance under field conditions, and aligning photonics vendors to exploit broader spectral windows. The most likely near‑term trajectory is selective, high‑value deployment—data‑center interconnects, trading links and specialized scientific or sensing applications—followed by gradual expansion as production, cost and interoperability metrics mature.
Hyperscalers and carriers planning network roadmaps should adopt pragmatic optimism: invest in pilots where the benefits are compelling, require independent validation and operational acceptance tests, and negotiate supplier commitments that reflect the realities of bringing a new fiber class from lab to long‑haul operations. The DNANF milestone is a major technical achievement—and its commercial significance will be determined by the quality of follow‑through engineering and standards work over the coming years.
Conclusion
DNANF hollow‑core fiber is not a marketing fantasy: it is an experimentally validated advance built on decades of hollow‑core physics and years of focused R&D. The technology promises lower loss, lower latency and more spectral headroom—advantages that can reshape how cloud and network operators design high‑value routes. The central question now is not whether DNANF works in the lab, but whether industry can translate spool‑level success into repeatable manufacturing, robust field operations and an interoperable ecosystem. The next 24–48 months will decide whether this physics milestone becomes a practical foundation for the network‑era demands of AI, real‑time services and hyperscale cloud.
Source: Optics.org Southampton achieves ‘record’ low signal loss with hollow fiber