The global hollow-core fiber market was valued at USD 0.18 billion in 2025. It is estimated at USD 0.34 billion in 2026 and is projected to reach USD 5.89 billion by 2036, growing at a CAGR of 33.0% over the 2026-2036 forecast period. The report gives a detailed assessment of the market, covering fiber technology, performance against conventional single-mode fiber, manufacturing and supply chains, standardization, intellectual property, pricing and total cost of ownership, the value chain, and the competitive landscape.
Hollow-core fiber carries light through a core filled with air or gas instead of solid glass. Light moves much faster in air and barely interacts with the glass structure, so these fibers offer lower latency, very low nonlinearity and dispersion, and, in the newest designs, lower loss than standard single-mode fiber. The study covers antiresonant fibers, among them the double-nested nodeless design (DNANF), and photonic bandgap fibers, together with cables for data center, metro, and backbone routes, the splices, connectors, and transition parts that link them to standard fiber, and dedicated test equipment. Fibers for laser delivery, sensing, and scientific instruments fall within scope, while solid-core fibers and cables do not.
Within a few years, hollow-core fiber has moved from the laboratory into live networks, led by Microsoft, which bought Lumenisity, a 2017 University of Southampton spin-off, in 2022, set up a fabrication site in the U.K., and has run its DNANF in several Azure regions since 2023. According to Microsoft, the fiber offers up to 47% faster data transmission and about 33% lower latency than single-mode fiber, and the company intends to install another 15,000 kilometers over a 24-month period. A 2025 paper in Nature Photonics by Microsoft Azure Fiber and Southampton researchers reported a loss of 0.091 dB per kilometer, below the roughly 0.14 dB per kilometer level that has limited silica fiber for about 40 years.
Growth is held back by the difficulty of drawing thin-membrane glass microstructures with precise geometry over long lengths at high yield, which keeps hollow-core fiber far more costly than single-mode fiber, and the record loss figure was a best-case result rather than a production average. Splicing tools, connectors, cable designs, and installation methods are only now arriving, and carriers usually want several qualified vendors before using a new fiber in backbone networks. Conventional fiber is cheap, close to its loss limit, and already installed at vast scale, and its saving of about 1.5 microseconds per kilometer matters only on latency-sensitive routes. Demand and know-how are also concentrated, since one hyperscaler owns the leading design and drives most near-term volume. Key challenges include low-loss joints to existing fiber and equipment, the lack of industry standards, proving decades of field reliability against bending, temperature, and moisture, and the need for new test methods and trained crews, as very low backscatter changes how faults and losses are measured.
Linking distributed AI campuses into larger virtual clusters is the largest near-term opportunity, as lower-latency fiber lets operators spread capacity across sites with available power. Trading firms in financial hubs such as London, New York, Chicago, and Frankfurt stand to gain, since a 100-kilometer route could save roughly 150 microseconds each way, and hollow-core fiber has already been used on trading routes. Mobile fronthaul links in 5G and 6G networks, laser delivery, gas sensing, fiber-optic gyroscopes, and quantum technologies add further demand. Key trends include production scale-up through partnerships with large glass and fiber makers, the rise of a multi-vendor ecosystem, such as Lightera's work with Nokia and Digital Realty and its 2025 splicing tools with Furukawa Electric, growing development and field trials in Asia, and new test and installation practices.
By fiber design, the report covers DNANF, other antiresonant fibers, and photonic bandgap fibers. DNANF is set to hold the largest share in 2026, as it is the design behind the record 0.091 dB per kilometer result and the fiber Microsoft runs across its Azure network, and its production is moving to Corning's North Carolina plants and Heraeus Covantis sites in Europe and the U.S. Other antiresonant fibers are projected to post the highest CAGR, as new entrants, including fiber makers in Asia and Europe, develop their own designs for telecom and data center use instead of relying on Microsoft's intellectual property.
By product, the study covers hollow-core fiber, hollow-core cables, connectivity & interconnection (splicing solutions and connectors & transition components), and test & measurement. Hollow-core cables are expected to lead in 2026, because fiber reaches buyers mostly in cabled form for data center interconnect and network routes, and cabling adds protection and installation value on top of the bare fiber. Connectivity & interconnection is projected to grow the fastest, as every hollow-core link must be joined to conventional fiber and equipment at both ends with low loss and low back-reflection, and splicing tools such as the 2025 Lightera and Furukawa Electric range are only now coming to market. Test & measurement gains as reflectometry methods are adapted to hollow-core fiber.
By application, the report covers data center interconnect, intra-data center & AI cluster interconnect, low-latency trading, long-haul & metro networks, mobile fronthaul & backhaul, and sensing & specialty uses such as laser delivery, gas sensing, gyroscopes, and quantum applications. Data center interconnect is set to account for the largest share in 2026, as hyperscalers are deploying the fiber first between data centers within cloud regions, where lower delay widens the area in which new sites can be built. Intra-data center & AI cluster interconnect is projected to post the highest CAGR, driven by the scale-up of AI clusters and the need for optical links with lower latency and lower power use. Low-latency trading is a small but high-value use where buyers accept a premium for each microsecond saved.
By end user, the study covers hyperscale cloud providers, telecommunications operators, financial services firms, data center operators, and research, defense & industrial users. Hyperscale cloud providers are expected to lead in 2026, since Microsoft owns its own routes and equipment and has committed to large deployments, which lets hyperscalers adopt the fiber well before standards are in place. Telecommunications operators are projected to grow the fastest, as standards mature, more suppliers qualify products, and use cases open in metro networks and mobile fronthaul. Financial services firms are early, premium buyers on specific routes between trading venues.
In 2026, North America is expected to hold the largest share of the global market, as it hosts the largest hyperscale deployments and the first industrial-scale production. Microsoft has run hollow-core routes in several Azure regions since 2023, Corning will make the fiber in North Carolina, and Heraeus Covantis will also produce in the U.S. Digital Realty has used Lightera's hollow-core fiber in data center settings, U.S. trading networks are likely early users, and Canada's cloud and financial markets offer further potential. Europe holds a significant share and is where modern hollow-core technology began, through the University of Southampton's work in the U.K.; Germany's Heraeus Covantis will produce from European sites, Lightera has also partnered with Finland's Nokia, specialty photonics firms in France and Denmark serve laser and sensing uses, and London and Frankfurt are candidate trading routes.
Asia-Pacific is expected to grow at the highest CAGR, supported by its lead in optical fiber manufacturing, fast data center growth, and active research and field trials. China makes more than half of the world's optical fiber, and Chinese fiber makers and operators have built hollow-core fibers and tested them in data center and backbone networks. Japan's Furukawa Electric, Fujikura, and Sumitomo Electric bring strong specialty fiber and connectivity capabilities. AI data center growth in Japan, South Korea, Singapore, India, and Australia adds demand. In Latin America, cloud providers are adding regions in Brazil, Mexico, and Chile, and Brazil's 2024 national AI plan foresees about BRL 23 billion of investment. The Middle East & Africa is expected to grow strongly from a small base, as Gulf AI campuses such as Stargate UAE, which starts with a 1-gigawatt cluster within a planned 5-gigawatt campus announced in May 2025, can be designed around hollow-core fiber from the start.
In this early-stage and highly concentrated market, Microsoft leads through its ownership of DNANF technology from the Lumenisity acquisition and its in-house and partner production, alongside major fiber and cable makers, specialty photonics firms, splicing and connectivity suppliers, test equipment companies, and network equipment vendors. Companies compete on loss and latency performance, yield and cost, reliability, compatibility with installed networks, intellectual property, and access to hyperscale buyers, mainly through production alliances, alternative designs, and new splicing and test tools. Microsoft's September 2025 manufacturing agreements with Corning and Heraeus Covantis mark the shift from specialist to industrial-scale production.
Key companies profiled in the report include Microsoft Corporation (Lumenisity) (U.S.), Corning Incorporated (U.S.), Heraeus Covantis (Germany), Lightera (U.S.), Yangtze Optical Fibre and Cable Joint Stock Limited Company (YOFC) (China), Furukawa Electric Co., Ltd. (Japan), Fujikura Ltd. (Japan), Sumitomo Electric Industries, Ltd. (Japan), Prysmian S.p.A. (Italy), Nokia Corporation (Finland), NKT Photonics (Denmark), GLOphotonics (France), VIAVI Solutions Inc. (U.S.), and EXFO Inc. (Canada).
Hollow-core fiber carries light through a core filled with air or gas instead of solid glass. Light moves much faster in air and barely interacts with the glass structure, so these fibers offer lower latency, very low nonlinearity and dispersion, and, in the newest designs, lower loss than standard single-mode fiber. The study covers antiresonant fibers, among them the double-nested nodeless design (DNANF), and photonic bandgap fibers, together with cables for data center, metro, and backbone routes, the splices, connectors, and transition parts that link them to standard fiber, and dedicated test equipment. Fibers for laser delivery, sensing, and scientific instruments fall within scope, while solid-core fibers and cables do not.
Within a few years, hollow-core fiber has moved from the laboratory into live networks, led by Microsoft, which bought Lumenisity, a 2017 University of Southampton spin-off, in 2022, set up a fabrication site in the U.K., and has run its DNANF in several Azure regions since 2023. According to Microsoft, the fiber offers up to 47% faster data transmission and about 33% lower latency than single-mode fiber, and the company intends to install another 15,000 kilometers over a 24-month period. A 2025 paper in Nature Photonics by Microsoft Azure Fiber and Southampton researchers reported a loss of 0.091 dB per kilometer, below the roughly 0.14 dB per kilometer level that has limited silica fiber for about 40 years.
Market Dynamics
The main driver is the need for faster, higher-capacity links in AI and cloud networks. Cloud regions combine several data centers that must be joined with very low delay, and the speed of light in glass caps how far apart they can be. Lower latency lets operators place sites farther apart within the same budget, opening up more land and power, which matters as AI training spreads across buildings and campuses. According to NVIDIA's chief financial officer, 2026 capital spending by the top five cloud providers and hyperscalers is approaching USD 700 billion, and the International Energy Agency expects data center power use to reach around 945 terawatt-hours by 2030, over twice the current level. Record performance strengthens the case further, since air-guided light also brings nonlinearity about three orders of magnitude below silica, which can stretch amplifier spacing and cut signal processing and energy needs. A committed anchor buyer adds confidence: Microsoft's September 2025 agreements with Corning and Heraeus Covantis give the industry its first industrial-scale manufacturing base.Growth is held back by the difficulty of drawing thin-membrane glass microstructures with precise geometry over long lengths at high yield, which keeps hollow-core fiber far more costly than single-mode fiber, and the record loss figure was a best-case result rather than a production average. Splicing tools, connectors, cable designs, and installation methods are only now arriving, and carriers usually want several qualified vendors before using a new fiber in backbone networks. Conventional fiber is cheap, close to its loss limit, and already installed at vast scale, and its saving of about 1.5 microseconds per kilometer matters only on latency-sensitive routes. Demand and know-how are also concentrated, since one hyperscaler owns the leading design and drives most near-term volume. Key challenges include low-loss joints to existing fiber and equipment, the lack of industry standards, proving decades of field reliability against bending, temperature, and moisture, and the need for new test methods and trained crews, as very low backscatter changes how faults and losses are measured.
Linking distributed AI campuses into larger virtual clusters is the largest near-term opportunity, as lower-latency fiber lets operators spread capacity across sites with available power. Trading firms in financial hubs such as London, New York, Chicago, and Frankfurt stand to gain, since a 100-kilometer route could save roughly 150 microseconds each way, and hollow-core fiber has already been used on trading routes. Mobile fronthaul links in 5G and 6G networks, laser delivery, gas sensing, fiber-optic gyroscopes, and quantum technologies add further demand. Key trends include production scale-up through partnerships with large glass and fiber makers, the rise of a multi-vendor ecosystem, such as Lightera's work with Nokia and Digital Realty and its 2025 splicing tools with Furukawa Electric, growing development and field trials in Asia, and new test and installation practices.
Segment Analysis
The report analyzes the market by fiber design, product, application, end user, and geography, helping stakeholders identify where adoption will start and how it will spread across networks and buyers.By fiber design, the report covers DNANF, other antiresonant fibers, and photonic bandgap fibers. DNANF is set to hold the largest share in 2026, as it is the design behind the record 0.091 dB per kilometer result and the fiber Microsoft runs across its Azure network, and its production is moving to Corning's North Carolina plants and Heraeus Covantis sites in Europe and the U.S. Other antiresonant fibers are projected to post the highest CAGR, as new entrants, including fiber makers in Asia and Europe, develop their own designs for telecom and data center use instead of relying on Microsoft's intellectual property.
By product, the study covers hollow-core fiber, hollow-core cables, connectivity & interconnection (splicing solutions and connectors & transition components), and test & measurement. Hollow-core cables are expected to lead in 2026, because fiber reaches buyers mostly in cabled form for data center interconnect and network routes, and cabling adds protection and installation value on top of the bare fiber. Connectivity & interconnection is projected to grow the fastest, as every hollow-core link must be joined to conventional fiber and equipment at both ends with low loss and low back-reflection, and splicing tools such as the 2025 Lightera and Furukawa Electric range are only now coming to market. Test & measurement gains as reflectometry methods are adapted to hollow-core fiber.
By application, the report covers data center interconnect, intra-data center & AI cluster interconnect, low-latency trading, long-haul & metro networks, mobile fronthaul & backhaul, and sensing & specialty uses such as laser delivery, gas sensing, gyroscopes, and quantum applications. Data center interconnect is set to account for the largest share in 2026, as hyperscalers are deploying the fiber first between data centers within cloud regions, where lower delay widens the area in which new sites can be built. Intra-data center & AI cluster interconnect is projected to post the highest CAGR, driven by the scale-up of AI clusters and the need for optical links with lower latency and lower power use. Low-latency trading is a small but high-value use where buyers accept a premium for each microsecond saved.
By end user, the study covers hyperscale cloud providers, telecommunications operators, financial services firms, data center operators, and research, defense & industrial users. Hyperscale cloud providers are expected to lead in 2026, since Microsoft owns its own routes and equipment and has committed to large deployments, which lets hyperscalers adopt the fiber well before standards are in place. Telecommunications operators are projected to grow the fastest, as standards mature, more suppliers qualify products, and use cases open in metro networks and mobile fronthaul. Financial services firms are early, premium buyers on specific routes between trading venues.
Regional Analysis
The report covers North America, Europe, Asia-Pacific, Latin America, and the Middle East & Africa, with analysis of 21 countries (with Finland assessed together with the other Nordic countries) and five rest-of-region groups. The regional analysis considers hyperscale and AI data center investment, fiber manufacturing capacity, research activity, financial trading hubs, and telecom network plans.In 2026, North America is expected to hold the largest share of the global market, as it hosts the largest hyperscale deployments and the first industrial-scale production. Microsoft has run hollow-core routes in several Azure regions since 2023, Corning will make the fiber in North Carolina, and Heraeus Covantis will also produce in the U.S. Digital Realty has used Lightera's hollow-core fiber in data center settings, U.S. trading networks are likely early users, and Canada's cloud and financial markets offer further potential. Europe holds a significant share and is where modern hollow-core technology began, through the University of Southampton's work in the U.K.; Germany's Heraeus Covantis will produce from European sites, Lightera has also partnered with Finland's Nokia, specialty photonics firms in France and Denmark serve laser and sensing uses, and London and Frankfurt are candidate trading routes.
Asia-Pacific is expected to grow at the highest CAGR, supported by its lead in optical fiber manufacturing, fast data center growth, and active research and field trials. China makes more than half of the world's optical fiber, and Chinese fiber makers and operators have built hollow-core fibers and tested them in data center and backbone networks. Japan's Furukawa Electric, Fujikura, and Sumitomo Electric bring strong specialty fiber and connectivity capabilities. AI data center growth in Japan, South Korea, Singapore, India, and Australia adds demand. In Latin America, cloud providers are adding regions in Brazil, Mexico, and Chile, and Brazil's 2024 national AI plan foresees about BRL 23 billion of investment. The Middle East & Africa is expected to grow strongly from a small base, as Gulf AI campuses such as Stargate UAE, which starts with a 1-gigawatt cluster within a planned 5-gigawatt campus announced in May 2025, can be designed around hollow-core fiber from the start.
Competitive Landscape
The report reviews key growth strategies, competitive benchmarking, the competitive dashboard, and 2025 market share/rank analysis by key player. It profiles hollow-core fiber developers and producers, fiber, cable, and connectivity companies, and specialty photonics and test companies, covering business overview, financial overview, product portfolio, strategic developments, and SWOT analysis.In this early-stage and highly concentrated market, Microsoft leads through its ownership of DNANF technology from the Lumenisity acquisition and its in-house and partner production, alongside major fiber and cable makers, specialty photonics firms, splicing and connectivity suppliers, test equipment companies, and network equipment vendors. Companies compete on loss and latency performance, yield and cost, reliability, compatibility with installed networks, intellectual property, and access to hyperscale buyers, mainly through production alliances, alternative designs, and new splicing and test tools. Microsoft's September 2025 manufacturing agreements with Corning and Heraeus Covantis mark the shift from specialist to industrial-scale production.
Key companies profiled in the report include Microsoft Corporation (Lumenisity) (U.S.), Corning Incorporated (U.S.), Heraeus Covantis (Germany), Lightera (U.S.), Yangtze Optical Fibre and Cable Joint Stock Limited Company (YOFC) (China), Furukawa Electric Co., Ltd. (Japan), Fujikura Ltd. (Japan), Sumitomo Electric Industries, Ltd. (Japan), Prysmian S.p.A. (Italy), Nokia Corporation (Finland), NKT Photonics (Denmark), GLOphotonics (France), VIAVI Solutions Inc. (U.S.), and EXFO Inc. (Canada).
How This Report Helps
- Provides reliable market size estimates and long-term forecasts for the global hollow-core fiber market through 2036.
- Explains how AI and cloud latency needs, record fiber performance, and hyperscaler commitments are shaping adoption.
- Assesses the impact of manufacturing yield, ecosystem gaps, conventional fiber economics, and concentrated intellectual property.
- Identifies high-growth opportunities across fiber design, product, application, end user, and region.
- Tracks production partnerships, alternative fiber designs, splicing and test tools, and standardization efforts.
- Benchmarks fiber developers, cable and connectivity makers, and test companies on technology and market position.
- Supports product planning, partnership, sourcing, and investment decisions for fiber makers, network operators, data center owners, and investors.
Key Questions Answered
- How large is the global hollow-core fiber market today, and how will it grow through 2036?
- How does hollow-core fiber compare with single-mode fiber on latency, loss, nonlinearity, and total cost of ownership?
- What are the main drivers, restraints, opportunities, challenges, and trends shaping the market?
- How quickly can manufacturing yield and volume improve as production moves to major fiber makers?
- Which fiber design, product, application, and end user segments will grow the fastest?
- When will telecom operators and trading firms adopt hollow-core fiber beyond hyperscale deployments?
- Which regions and countries offer the most attractive opportunities, and what role will Asian fiber makers play?
- Who are the leading companies in the market, and how do they compare?
- What partnerships, product launches, and deployments are shaping competition?
Table of Contents
1. Introduction
2. Research Methodology
4. Market Overview
5. Hollow-Core Fiber Market, by Fiber Design
6. Hollow-Core Fiber Market, by Product
7. Hollow-Core Fiber Market, by Application
8. Hollow-Core Fiber Market, by End User
9. Hollow-Core Fiber Market, by Geography
10. Competitive Landscape
11. Company Profiles
12. Appendix
LIST OF TABLES
LIST OF FIGURES
Companies Mentioned
- Microsoft Corporation
- Corning Incorporated
- Heraeus Covantis
- Lightera
- Yangtze Optical Fibre and Cable Joint Stock Limited Company
- Furukawa Electric Co., Ltd.
- Fujikura Ltd.
- Sumitomo Electric Industries, Ltd.
- Prysmian S.p.A.
- Nokia Corporation
- NKT Photonics
- GLOphotonics
- VIAVI Solutions Inc.
- EXFO Inc.
