Triple Inflection Points Converge: Great Wall Fund's Han Lin Bullish on Fiber Optic Investment Opportunities

Deep News
04/27

The rapid advancement of artificial intelligence (AI) technology is fueling a surge in AI computing power, sweeping through the A-share market and positioning the entire optical communication industry chain as a high-potential investment theme for 2026. Fiber optic cables, optical modules, optical chips, and optical components are experiencing frequent rotational gains and successive upward movements. The fiber optic industry, serving as the "physical arteries" of computing power networks, is undergoing a significant revaluation—transitioning from a traditional telecommunications consumable to essential computing infrastructure. Resonating with three key inflection points—explosive demand, inelastic supply, and price recovery—the sector is entering a golden window characterized by rising volumes and prices, profit restoration, and improved industry structure. Han Lin, Portfolio Manager of the Great Wall Digital Economy Mixed Fund, recently shared his professional insights on the investment value of the fiber optic industry in a media interview.

**Question 1: Why focus on the fiber optic industry starting in 2026?** Han Lin, Portfolio Manager of the Great Wall Digital Economy Mixed Fund: We are firmly incorporating the fiber optic industry into our key focus areas for 2026. The core reason is that this year marks a historic shift for the industry—from traditional telecommunications cycles to the rigid demand driven by AI computing power. Concurrently, a resonance is occurring across three inflection points: supply-demand dynamics, pricing structures, and profit expectations, creating a rare strategic window for allocation. From the demand perspective, the computing revolution brought by generative AI is fundamentally rewriting the scale and structure of fiber optic demand. Data from a leading global fiber optic cable manufacturer indicates that a 32-GPU server requires four times the fiber of a traditional cloud server, and a 72-GPU node quadruples that amount again. Internal data center interconnects and Data Center Interconnect (DCI) have become the primary growth engines. According to the latest CRU data, global data center demand for fiber optic cable surged 75.9% year-over-year in 2025. The proportion of AI-related fiber demand, which was below 5% in 2024, is expected to rapidly increase to 30%-35% by 2027. More optimistic industry views suggest this share could exceed 50% as global hyperscale AI data centers are densely deployed. Specifically, demand for G.657.A2 bend-insensitive fiber has exploded due to high-density cabling and narrow-space routing requirements in data centers. Its price has already risen to over 260 yuan per fiber-kilometer, making it a high-end, essential product growing significantly faster than traditional fiber. On the supply side, there are strong rigid constraints. Optical preform, which accounts for 70% of total fiber cost, represents the core technological and capacity barrier in the industry, with expansion cycles lasting 18-24 months. Currently, major global manufacturers are operating at full capacity, actively prioritizing production towards higher-margin specialty fibers. Supply of standard G.652.D fiber continues to shrink. Furthermore, export controls on germanium tetrachloride and significant price increases for helium have further tightened the supply of upstream raw materials. Clear signals of price reversal have emerged. G.652.D fiber prices have rapidly climbed from below 20 yuan per fiber-kilometer in early 2025 to a range of 100-120 yuan. Bulk procurement prices from operators like Tianjin Telecom and Heilongjiang Telecom have been significantly raised, indicating that telecom operators are gradually accepting and acknowledging the rationale for price increases. The industry has decisively moved beyond years of low-price competition, allowing profit elasticity to be fully unleashed. Additionally, several major domestic leaders hold top global market shares, with overseas orders growing steadily. Recovery in North American FTTH demand and accelerated computing infrastructure build-out in the Middle East are opening further incremental overseas opportunities. Moreover, global fiber consumption for drone applications reached approximately 50-80 million fiber-kilometers in 2025, creating additional阶段性 supply pressure. With multiple positive factors converging, 2026 is poised to be the time to position for the triple benefits of volume-price growth, profit recovery, and structural optimization in the fiber optic industry, leading us to decisively increase our focus.

**Question 2: How does the investment logic for fiber optics change with its shift from "telecom cycles" to "computing necessity" driven by AI? How does the corresponding valuation system evolve?** Han Lin, Portfolio Manager of the Great Wall Digital Economy Mixed Fund: AI's reshaping of the fiber optic industry represents a fundamental shift in its underlying logic—from a cyclical telecommunications consumable to growth-oriented computing infrastructure. The sector's attributes are evolving from "cyclical stocks" towards "cyclical growth stocks," with potential to further transition into "growth stocks." This has led to fundamental and irreversible changes in both investment logic and valuation frameworks. Traditionally, the fiber optic industry was entirely dependent on telecommunications investment cycles, with demand heavily driven by operator bulk procurement for projects like 5G deployment and FTTH coverage. The industry exhibited a characteristic 7-year cycle of concentrated demand surges, rapid capacity overhangs, and vicious price competition. Investment focus was solely on identifying cyclical reversal points and short-term profit recovery. Valuations consistently hovered at levels typical for cyclical stocks in traditional manufacturing, with relatively low PE midpoints, as the market priced only for cycles and assigned no growth premium. In the AI computing era, the nature of demand has fundamentally changed. It has shifted from passive, cyclical, and homogeneous telecom demand to active, rigid, continuous, and premium growth demand. Three key scenarios—interconnects within AI data centers, DCI for linking data centers, and Scale-Up intra-cabinet optical interconnects—are upgrading fiber from a basic communication material to the core foundation of computing networks. Notably, AI is powerfully unlocking demand for new types of optical fiber, accelerating the industry's structural upgrade towards higher-end products. Consequently, the investment logic has completely switched. The focus is no longer on tracking operator procurement cycles and price negotiations but on monitoring growth indicators like AI server shipments, data center construction progress, penetration rates of high-end fibers, and the finalization of overseas orders. Leading companies, leveraging three core barriers—in-house preform capability, advanced technology reserves, and global footprint—are seeing their earnings transition from cyclical volatility to stable, sustained growth. The valuation system has been simultaneously reconstructed. The industry is shedding its cyclical label and is being benchmarked against core assets within the computing power产业链. The PE midpoint is expected to rise significantly. Leading companies with high proportions of premium products, ample overseas orders, and deep technological moats may command additional valuation premiums. The profit model has also transformed from competing on price and relying on high volume with low margins to benefiting from simultaneous volume and price increases, product mix upgrades, and global expansion. Leading enterprises with over 80% self-sufficiency in optical preform are seeing substantially improved profit retention rates. The共振 between accelerating earnings growth and expanding valuations creates the potential for a "Davis Double Play." This represents the most essential and core distinction between the current market dynamic and any previous rallies driven solely by telecom cycles.

**Question 3: Within the industry chain—encompassing upstream core raw materials (preform, quartz), midstream fiber and cable manufacturing, downstream core optical communication components, AI data centers, and overseas markets—which segments present the most promising growth opportunities, and how long might this growth last?** Han Lin, Portfolio Manager of the Great Wall Digital Economy Mixed Fund: Across the entire fiber optic communication industry chain, we are relatively optimistic about three segments: upstream optical preform and quartz raw materials, midstream high-end fiber and cable, and downstream optical chips and modules. These segments possess strong potential and relatively long growth runways, with overall high景气 expected to last until the end of 2027, and possibly extending to 2030 for some high-end technological segments. Upstream optical preform and quartz represent one of the most critical barrier segments in the entire chain. Preform accounts for 70% of fiber cost, features extremely high technological barriers, and has very long expansion cycles. Major domestic players boast preform self-sufficiency rates exceeding 80%, firmly holding pricing power and positioning them as the primary beneficiaries of industry-wide price increases. Quartz, a core raw material for fiber production, benefits from supply rigidity that continues to support price increases. The potential of this upstream segment spans the entire景气 cycle, with growth likely sustainable for 2-3 years. In the midstream fiber and cable segment, we focus on high-end niches. Standard G.652.D fiber is benefiting from supply-demand mismatches and experiencing price recovery. However, high-end fibers like G.657.A2 and G.654.E are directly tied to the rigid demands of AI data centers and are potentially irreplaceable in high-density interconnect and long-distance, low-loss transmission scenarios, offering significant volume-price elasticity. The growth duration for this segment is largely synchronized with the pace of global computing infrastructure build-out, likely lasting around 2 years. Downstream optical chips and modules are key conduits for transmitting computing power demand. High-speed 800G and 1.6T optical modules are rapidly scaling within AI training and inference clusters. The localization of high-end optical chips is accelerating. Silicon photonics chips are gaining traction as replacement opportunities due to advantages like high integration, low power consumption, and cost efficiency. This segment has strong growth attributes and a wide growth space, potentially sustainable for 3-5 years. AI data centers are the ultimate source of demand for the entire chain. Overseas tech giants have already locked in large fiber procurement orders extending to 2030. Domestically, the ongoing development of the integrated "East Data West Computing" computing network provides a solid demand foundation. In overseas markets, recovering FTTH demand in North America, computing infrastructure upgrades in Europe, and accelerated digital construction in Southeast Asia contribute stable incremental demand. The continued increase in global market share for domestic leaders could form a second growth curve. In summary, these three core segments form a pattern of gradient growth and successive advancement: stable growth upstream (preform), strong elasticity midstream (high-end fiber), and high growth downstream (chips/modules). They support and drive each other, collectively sustaining high industry景气. This presents a rare high-potential direction within the current A-share market landscape, characterized by no significant weaknesses or断层.

免責聲明:投資有風險,本文並非投資建議,以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請,作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考,不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證,投資者應自行研究並在投資前尋求專業建議。

熱議股票

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10