FTTX Fiber Is Manufactured Globally

 

The Preform Problem Nobody Talks About

 

Here's what the glossy industry reports won't emphasize: preform manufacturing is where the real bottleneck lives.

A preform is essentially a fat glass rod-maybe 200mm in diameter and three meters long-from which manufacturers draw hundreds of kilometers of finished fiber. The chemistry involved requires silicon tetrachloride and germanium tetrachloride reacting with oxygen at temperatures that would make most industrial processes look quaint. We're talking 1900°C furnaces, deposition rates measured in grams per minute, and purity requirements where parts-per-billion contamination levels actually matter.

China produces something like 90% of the world's germanium, which gives them structural cost advantages that competitors struggle to match. I've seen procurement managers at European telecoms quietly admit they've basically given up trying to source non-Chinese preforms at competitive prices. The math just doesn't work.

MCVD (Modified Chemical Vapor Deposition) remains the workhorse process-layers of doped silica soot deposited inside rotating quartz tubes, then collapsed into solid rods through intense heating. Bell Labs pioneered this in the 1970s using Heraeus fused quartz tubes, and the fundamental approach hasn't changed as much as you'd expect. OVD and VAD variations exist, each with different tradeoffs between deposition efficiency and refractive index control, but MCVD dominates commercial production.

The Japanese manufacturers-Sumitomo, Fujikura, Shin-Etsu-still command respect for preform quality. Their stuff runs maybe 15-20% premium over Chinese equivalents, but for submarine cables or specialty applications where you absolutely cannot have attenuation problems at kilometer 47, people pay it.

 

 

Where the Fiber Actually Comes From

 

YOFC (Yangtze Optical Fibre and Cable) operates out of Wuhan and they're the world's second-largest optical fiber manufacturer after Corning. Their production capacity exceeds 100 million fiber kilometers annually. That number is difficult to contextualize until you realize it's enough to wrap around Earth's equator roughly 2,500 times. Every year.

Hengtong and ZTT round out the Chinese giants. Hengtong has been particularly aggressive about vertical integration-they've acquired facilities across the supply chain and now produce everything from preforms to finished cable assemblies to submarine systems. ZTT claims their optical cable factory alone covers 40,000 square meters.

The concentration bothers some telecom infrastructure planners. Not for political reasons necessarily, but because supply chain resilience becomes genuinely questionable when 50+ million core kilometers of annual production sits within a few hundred miles of each other in eastern China. The 2020-2021 shipping disruptions exposed how fragile these logistics chains can be.

Prysmian operates 104 plants across 50+ countries, which sounds diversified until you dig into where the optical fiber specifically gets made versus where they're just doing cable assembly and termination. The high-value manufacturing steps cluster more than the plant counts suggest.

 

Bend-Insensitive Fiber Changed Everything (Mostly)

 

The ITU-T G.657 standard emerged to solve a problem installers had been complaining about for years: conventional single-mode fiber hated tight bends.

Standard G.652 fiber needed 30mm bend radii to keep macrobending losses acceptable. That's fine for outdoor trunk cables running through conduit, but completely impractical for the last hundred meters of an FTTH deployment where the cable has to navigate around door frames, through wall corners, and into cramped splice enclosures.

G.657.A1 pushed the minimum bend radius down to 10mm. G.657.A2 and B2 variants hit 7.5mm. The B3 category-which I still think is slightly insane-allows 5mm bends with less than 0.15 dB loss at 1550nm.

The engineering trick involves modifying the refractive index profile to tighten the mode field diameter and increase numerical aperture. There are tradeoffs-some B3 fibers show elevated chromatic dispersion or splice compatibility issues with legacy G.652 infrastructure-but for access network deployments where you're running fiber to apartment units with whoever-knows-what installation practices, the bend tolerance matters more than theoretical performance optimizations.

I watched a contractor staple-gun FTTH drop cable along a baseboard last year. The cable survived. The traditional fiber from five years earlier absolutely would not have.

 

 

The FTTX Alphabet Soup

 

FTTH. FTTP. FTTB. FTTC. FTTN. The acronyms multiply because the fiber termination point keeps shifting based on economics and existing infrastructure.

FTTH means fiber runs directly into the residence-an optical network terminal sits inside somebody's home, converting photons to electrons for their router. This is the expensive option. Digging up streets, pulling cable through existing conduit if you're lucky, mounting equipment in every single dwelling unit. European deployments are seeing €500-800 per premises passed in urban areas; rural costs can hit multiples of that.

FTTB stops at the building's basement or communications room, then uses existing copper or ethernet cabling for the final distribution. Apartment complexes love this approach because you're not running new cable to every individual unit.

FTTN terminates at a neighborhood cabinet-the "node"-where copper completes the last mile. Cheaper to deploy, significantly worse performance over distance. The copper segment limits you in ways that make fiber's bandwidth advantages largely theoretical.

The distinctions matter less than the marketing suggests. What matters is where the optical-to-electrical conversion happens, and how much copper sits between that conversion point and the end user. More copper equals more attenuation, more electrical interference susceptibility, and lower achievable data rates. There's no magic here.

 

Quality Varies More Than Anyone Admits

 

Chinese manufacturers have earned a reputation for price competitiveness. Whether they've earned equivalent reputation for quality depends heavily on which supplier you're discussing and who's asking.

The major players-YOFC, Hengtong, ZTT, Futong-export to global telecoms and generally meet international certifications. ISO 9001, TIA-568, IEC 60794 compliance. Their cables test within spec. The factories look modern.

But the tier-two and tier-three suppliers? The ones quoting prices 40% below established manufacturers? That's where procurement gets interesting. I've heard stories-secondhand, admittedly-of fiber with suspiciously high attenuation, water-blocking compounds that degraded within two years, and cable jackets that cracked in mild UV exposure.

The testing burden shifts to buyers in ways that aren't always obvious. A reputable manufacturer's quality control catches defects before shipment. Cheaper sources might require incoming inspection on every spool-OTDR traces, tensile testing, jacket thickness verification-which erases some of the cost savings.

None of this is unique to Chinese manufacturing. I've seen domestically-produced cables fail in spectacular fashion too. But when you're optimizing purely for purchase price on a commodity product, the quality variance increases.

 

 

Installation Realities

 

The fiber itself is almost never the problem.

Installation practices cause probably 80% of the performance issues I've seen in deployed FTTH networks. Macrobends where the cable got kinked during pull-through. Connectors contaminated with dust because somebody didn't cap them. Fusion splices with excessive loss because the cleave was sloppy. Cable jackets damaged by overeager staple guns.

Bend-insensitive fiber helps. Pre-connectorized solutions reduce field termination errors. But fundamentally, the gap between laboratory specifications and real-world performance comes down to whoever's actually handling the cable.

The manufacturers know this. It's why companies like OFS offer professional services for turnkey deployments-they've learned that selling excellent fiber means nothing if the installation ruins it. CommScope packages test equipment with their FTTH products. The ecosystem has evolved to acknowledge that the last few meters of installation represent the highest-risk portion of the entire network.

 

Where This Is Heading

 

Hollow-core fiber remains mostly experimental but the physics are promising-guiding light through air instead of glass reduces latency and theoretically eliminates nonlinear effects that limit power handling. Prysmian has published work on it. Don't expect commercial FTTH deployments anytime soon.

Multi-core fibers pack multiple light-carrying cores into a single cladding, boosting capacity per cable without proportional size increases. Space-division multiplexing could matter for data center interconnects where duct space is constrained and bandwidth demands keep climbing.

200-micron reduced-diameter fibers are already shipping for high-density applications. Standard fiber is 250 microns with coating; trimming that allows more fibers per tube, more tubes per cable, higher counts per duct. The mechanical reliability at smaller diameters took years to validate, but the industry seems comfortable with it now.

The manufacturing geography probably won't shift dramatically. The capital investment required for preform production-hundreds of millions of dollars for a competitive facility-creates enormous barriers to entry. Companies that missed the buildout wave in the 2000s and 2010s aren't catching up easily.

If anything, I'd expect continued consolidation. Prysmian acquired General Cable. Corning keeps expanding capacity. Chinese manufacturers keep pushing into African and Southeast Asian markets. The number of truly independent, globally-competitive fiber manufacturers might actually shrink over the next decade even as total production volume climbs.

 

What Actually Matters

 

The fiber matters less than people think. Specifications have largely converged-any name-brand G.657.A2 fiber performs similarly to any other within the same standard category. The differentiation happens in cable construction, connector quality, and installation support.

If you're deploying FTTH infrastructure and agonizing over which fiber manufacturer to source from, you're probably optimizing the wrong variable. Spend that energy on installation training, test equipment, and post-deployment verification instead.

The global manufacturing base means competitive pricing across geographies. Use that to your advantage. But remember that a $50,000 savings on fiber procurement means nothing if poor installation practices create a network that requires $200,000 in remediation within three years.

The glass itself has been essentially commoditized. Everything around it still matters enormously.