Surging by 650%, explosive orders worldwide—what makes this optical fiber bull market different?
```
The optical fiber industry is undergoing an unprecedented supply and demand restructuring.
On April 20, the A-share optical fiber concept exploded collectively again—Tongding Internet hit the daily limit, YOFC, Zhongtian Technology, Hengtong Optic-Electric, and OpTelecom followed suit. Analysts point out that the market’s agitation is not groundless; the booming market is backed by strong fundamentals.
CCTV Finance latest report reveals a shocking set of data: G.657.A2 optical fiber prices have surged from 32 yuan per core kilometer last year to 240 yuan, an increase of 650%; a certain optical fiber company in Jiangsu saw production and sales in the first quarter increase by nearly 5 times year-on-year, with orders scheduled through the first quarter of next year.
However, the truly thought-provoking question is: What is fundamentally different about this round of optical fiber bull market compared to the past?
Guosheng Securities provided a clear answer in a research report in mid-March: this round of optical fiber and cable price increases are not a periodic recovery in the traditional operator cycle, but a "supply-demand gap driven price hike" jointly fueled by AI data centers and military drones—the buyer's market is turning into a seller's market, and the pricing system is being reconstructed.
According to Guosheng Securities’ estimates, if the average price of loose fiber rebounds to 90 yuan per core kilometer in 2026, optical fiber and cable business alone can bring 7.7 billion yuan and 5.3 billion yuan net profit elasticity (pre-tax) to YOFC and Hengtong Optic-Electric respectively; the supply-demand gap is expected to reach 6% in 2026 and further expand to 15% in 2027, and with the hard constraint of an optical preform expansion cycle lasting 18-24 months, there is no short-term solution.
Demand Restructuring: AI and Drones Ignite “Crowding-Out” Growth
Guosheng Securities’ research report points out that the biggest difference in this bull market is a fundamental change in the demand function. In the past, optical fibers were highly dependent on 5G and broadband construction, but now, AI and drones have become new core drivers that cannot be postponed and are insensitive to price.
First, AI data centers (AIDC) have completely changed the demand for optical fiber.
Traditional data centers process "north-south traffic," which is user access to servers. To save costs, they use a network architecture with 3:1 or even higher convergence ratio, multiple servers share an uplink, and each cabinet uses about 40-80 fiber cores.
AI data centers, in contrast, process "east-west traffic"—parameter synchronization between GPUs. In large GPU clusters, any latency causes the bottleneck effect, so a 1:1 non-blocking architecture is required and each GPU needs a dedicated high-speed optical fiber channel.
Taking the Nvidia DGX H100/H200 SuperPOD cluster architecture as an example, a single server requires as many as 32 optical fibers internally, and with the interconnection needs of the two-layer Spine-Leaf switch architecture, the optical fiber consumption per cabinet is more than 5-10 times higher than traditional cabinets, reaching 800-1000 fiber cores per cabinet.
The report says, Peng Chuyu, manager of Fiber R&D at Fiberhome, confirmed this: "In conventional data centers, data centers with less than 10,000 GPUs two years ago might have used only a few thousand kilometers of fiber, but now with enhanced computing power, demand has grown 5-10 times."
According to CRU data, AIDC optical fiber and cable demand as a proportion of global demand is expected to surge from 5% in 2024 to 30% in 2027, making data centers the core growth pole for the fiber market instead of telecom operators.
Guosheng Securities used both Capex mapping and optical module connection methods for cross-verification, and the two approaches are highly consistent: By 2026, the combined optical fiber and cable demand for data centers in China and North America will exceed 50 million core kilometers, accounting for over 10% of the global traditional market (base 500 million core kilometers).
Besides AI, military drones represent the most easily overlooked yet crucial incremental demand in this round.
Optical fiber has advantages in drone guidance that 4G/5G cellular and satellite communications cannot replace: physical connection, not subject to electromagnetic interference, strong resistance to electronic warfare. Drones drag a spool of fiber behind them, primarily applied in FPV attack drones, reconnaissance drones, and special tactical missions, with G.657A2 being the main model used.
The key is its consumable nature: a single drone typically consumes 10-20 kilometers of fiber, and after completing the mission it cannot be recycled, turning fiber from infrastructure into a high-frequency consumable item.
According to Guosheng Securities estimates, current drone-related fiber demand is about 50 million core kilometers per year. With continued conflict and increased penetration of fiber-guided models, future demand may exceed 80 million core kilometers per year. Even with eased conflict, annual demand will remain around 30 million core kilometers.
Guosheng Securities states that, considering AI Data Centers and drone demand together, new demand already exceeds 20% of the traditional market, which is the direct reason for the global shortage of high-quality fiber.
Supply Locked In: Optical Preform Capacity as the Absolute “Hard Constraint”
Demand explosion is only half the story; rigid supply constraints are the core reason for price elasticity far exceeding historical levels in this round.
Global optical fiber and cable capacity is highly concentrated: China accounts for over 60%, the US is centered on Corning, and Japan is represented by Fujikura. Current nominal global capacity is about 550-600 million core kilometers per year, and overseas expansion is extremely restrained.
What really limits industry-wide supply is the upstream optical fiber preform (optical preform) segment. Optical preforms account for about 70% of profit distribution in the industry chain. Scaling production faces uncompressible "temporal rigidity":
It involves the complex VAD/OVD chemical deposition process, strict approvals for hazardous materials, special cleanroom construction, and long yield ramp-ups; even with mature technology, expansion needs 18-24 months.
This means that even if manufacturers start expansion immediately, new capacity will not translate into effective supply until 2027.
There is also another “structural mismatch” on the supply side: leading manufacturers preferentially switch limited capacity to high value-added models like G.657.A1/A2 and G.654.E, causing crowd-out effects for ordinary fiber G.652.D used in traditional broadband. Difficulty and high costs in switching between models mean it’s hard to flexibly adjust capacity in a short time, exacerbating the supply-demand gap.
In addition, tariff barriers and anti-dumping policies fragment the global fiber market, so "nominal capacity" cannot be turned into effective "circulating capacity", reducing global fiber and cable liquidity.
Profit Elasticity: Every 10 Yuan Rise in Price Adds 200-1000 Million Yuan Net Profit for Leading Manufacturers
With supply limited and costs rigid, price rises almost fully translate into profit without additional input—this is the essential difference between this cycle and historical ones.
The fiber business has typical high operating leverage: leading companies have high self-sufficiency in preforms, and within the existing capacity framework, the rod-to-fiber-to-cable process doesn’t involve large-scale variable cost input. The revenue increment from price rises, after small increases in energy, materials, labor, almost entirely sink into profit.
Guosheng Securities calculates that every 10 yuan per core kilometer rise in loose fiber price gives leading manufacturers net profit elasticity of 200-1000 million yuan.
If by 2026, the average loose fiber price returns to 90 yuan per core km (referencing the 2018 prosperity peak and current spot price overseas), fiber and cable business alone can bring YOFC and Hengtong Optic-Electric 7.7 billion yuan and 5.3 billion yuan net profit elasticity (pre-tax).
The Biggest Difference Between This Bull Market and History: From Domestic Cycle to Global Resonance
Guosheng Securities research clearly points out that historical fiber cable price cycles were basically decided by domestic operators’ demand—2010 benefited from 3G construction, 2015-2017 benefited from China Mobile FTTH construction, all were short cycles driven domestically. This cycle is fundamentally different from the past in three dimensions:
First, the demand source has shifted from domestic to global resonance.
Overseas CSPs (cloud service providers) have a strong willingness to secure future fiber supply in advance, forming a new “long-term contract locked price” normal. Fujikura signed mid- and long-term supply agreements with US customers; Corning and Meta reached multi-year supply agreements worth up to 6 billion USD, and are simultaneously expanding. As of early 2026, some fiber delivery cycles have exceeded 60 weeks. Domestic procurement prices will follow global supply and demand, not be independently priced.Second, demand characteristics have shifted from postponable to price-insensitive.
AI compute expansion is strongly correlated with fiber demand and cannot be postponed; drone fiber is consumable and unrecyclable, with obvious continued restocking. Customers are not price sensitive, prioritizing delivery security and specifications.Third, supply constraints are shifting from cyclical to structural.
After the last industry washout, manufacturers are restrained in expanding, and existing capacity has become almost fixed stock. The 18-24 month preform expansion cycle means supply cannot quickly respond to short-term price signals, and price elasticity will be significantly greater than history.
Additionally, Guosheng Securities points out that after the advent of various AI Agents, token usage is surging, which may further spur demand for traditional telecom traffic and thereby boost traditional fiber and cable demand and pricing, forming a dual price-raising logic of "AI demand crowds out high-end capacity → ordinary fiber also faces shortage".
Risk Disclaimer and TermsThe market involves risks, and investment requires caution. This article does not constitute personal investment advice, nor does it take into account the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views or conclusions in this article suit their particular situation. Investing based on this article is at your own responsibility. ```