2026-11-16 22:06
Introduction: Hollow-core fiber bets made a decade ago have perfectly positioned companies to ride the AI compute wave. The surge in AI compute demand is reigniting the optical fiber and cable industry, once thought to have hit growth saturation, ushering in a new upcycle.
The AI compute demand continues to intensify, a sector previously considered to have reached its growth ceiling, now entering a new phase of market expansion.
Changfei Fiber, headquartered in Wuhan's Optics Valley, has emerged as a standout player in this AI-driven rally. In the first three quarters of this year, Changfei Fiber achieved sales revenue of 19.1 billion RMB, a year-on-year increase of 86%. Since the beginning of the year, the company’s stock price has surged by over 400%.
This success was not sudden luck. At the "Cross-Cycle Power" seminar held at the "Optics Valley Seven Stars" on November 13, Nie Lei, Vice President of Changfei Fiber, revealed to Interface News that the company had begun developing hollow-core fiber and multi-core fiber for AI intelligent computing centers as far back as ten years ago. "At the time, we didn’t know these would be used in AI; some products were developed but might sit idle for ten or twenty years."
In reality, prior to this period, the optical fiber and cable industry had endured prolonged stagnation. After the peak cycle of 5G infrastructure deployment, the sector entered a deep trough marked by overcapacity and aggressive procurement price wars, compressing profitability. In 2024, Changfei Fiber recorded total revenue of only 12.197 billion RMB, with parent-company net profit reaching just 676 million RMB.
Amid structural shifts in demand, Nie Lei summarized Changfei Fiber’s cross-cycle resilience as: “Do what needs to be done, one step at a time. When the next wave of light arrives, we’ll keep moving forward.”
Changfei Fiber’s journey mirrors the broader evolution of Wuhan’s Optics Valley. Once ridiculed as “Crying Optics” during the steep decline of the optoelectronics industry, the region has since diversified beyond pure optical communications, building a robust industrial cluster encompassing optical communications, lasers, semiconductors, and smart manufacturing.
In this current AI cycle, the Optics Valley has nurtured a new generation of enterprises—Yangtze Memory Technologies, FiberHome Communications, Hua Gong Technology, Opticstar Technology, Gaode Infrared, and China CITIC Telecom Mobile—many of which are rooted in traditional industries but are now unlocking new growth avenues in the emerging technology wave.
Yet this is not the endgame. Facing the transformative forces of AI, Wuhan’s Optics Valley continues seeking additional growth drivers, striving to help more traditional industries find their place in the next technological revolution.
Changfei Fiber’s experience is not an isolated case. Across Wuhan’s Optics Valley, firms such as FiberHome Communications, Opticstar Technology, and Hua Gong Technology have also proactively positioned themselves across various technical frontiers, anticipating the next wave of industrial demand.
The transformation within the optical communications sector is particularly evident. Traditionally, optical fiber and cable served communication network infrastructure; today, with the scaling of AI compute clusters and data centers, applications are expanding into new domains. Yang Yong, Vice President of FiberHome Communications, noted that beyond conventional cabling, new opportunities are arising from data center interconnectivity, 10G access, and automotive optical communications. FiberHome has not only broken records in hollow-core fiber performance testing but also launched ultra-high-core-count cables and automotive optical communication solutions. Optical communications are now extending from terrestrial networks into vehicle interiors and compute clusters.
Meanwhile, AI compute infrastructure is driving the evolution of communication equipment toward higher performance. FiberHome has adopted a full-stack strategy spanning DCI, switches, and server systems. Its 51.2T NPO switch core components are now domestically produced, while its subsidiary, Yangtze Computing, has entered the top tier in procurement tenders among operators and financial institutions.
For Opticstar Technology, technological cycles present familiar opportunities. Over the past five decades, the company has navigated multiple industry cycles—from long-wavelength lasers and WDM components, to 5G optical modules, and now AI optical interconnects. Each shift in demand has triggered upgrades in core components and product architecture. At this year’s OFC conference, Opticstar unveiled the world’s first 6.4T NPO product; simultaneously, its Phase III bonded zone project has commenced, focusing on high-end components and modules essential for AI optical interconnects.
Hua Gong Technology has pushed its R&D investment even further upstream. Over the past eight years, the company has consistently invested in silicon photonics chip development, launching silicon photonic modules powered by self-developed chips into the global top tier. In the first three quarters of this year, its laser equipment exports grew by 50%, while optical module sales in North America and Southeast Asia surged nearly fourfold. Traditional laser equipment and emerging optical communications businesses are forming a new growth synergy.
Beneath these strategic moves lies a deeper understanding of technology cycles. Yang Yong emphasizes that next-generation technologies must be invested in before they turn a profit. “Current revenue comes from legacy products; next year’s competitiveness stems from products currently under development; and where we stand five years from now depends entirely on projects initiated today.”
This insight explains why traditional firms in the Optics Valley have successfully captured the AI wave. Whether it’s Changfei Fiber, Opticstar, or Hua Gong, their investments in hollow-core fiber, optical modules, and silicon photonics chips were not specifically designed for AI—but found critical applications when AI demand emerged.
This underscores a fundamental truth: whoever begins accumulating capabilities during periods when the market is still immature will be best positioned to seize growth opportunities when the next cycle arrives.
AI has opened new growth horizons for traditional industries in Optics Valley, yet when viewed through the lens of global competition, the region still lags significantly behind the world’s leading innovation hubs.
During the seminar, Professor Zhang Jianhua, Dean of the School of Economics at Huazhong University of Science and Technology, compared the “Optics Valley Seven Stars” with the “Silicon Valley Seven Giants.” The latter’s combined market cap stands at $17.4 trillion—approximately 60 times that of the former. In his view, market capitalization is merely a surface indicator; the deeper gap lies in industry influence and control over supply chains.
While Optics Valley has cultivated several globally competitive enterprises, it still lacks substantial leadership in industry standards, technical architectures, and interface definitions. Unlike Silicon Valley, which features platform-based giants that connect entire ecosystems, Optics Valley’s development remains heavily reliant on industrial parks and enterprise clustering. There remain too few platform enterprises capable of creating novel applications and defining new use cases across diverse sectors.
Zhang Jianhua identifies the gap in four dimensions: scale, domain and scenario coverage, organizational form, and capital access. Hubei boasts top-tier scientific and educational resources nationwide, but key challenges persist: how to transition research outcomes from labs to markets, how to build platform enterprises with strong industrial leverage, and how to attract capital willing to fund long-term technological innovation.
This implies that Optics Valley’s next stage of competition is not merely about nurturing more firms or expanding industrial scale—it requires filling critical gaps in the innovation chain.
According to Zhang Jianhua, the first priority is addressing structural gaps by cultivating platform leaders capable of integrating industrial resources and setting technical standards. Second is bridging the technology transfer gap by establishing a seamless pipeline from early-stage R&D and pilot validation to large-scale industrialization. Third is closing the capital gap through financial innovation, ensuring that tech firms at every stage receive funding aligned with their development lifecycle.
Enterprises are already recalibrating their strategies toward higher ambitions.
Changfei Fiber, for instance, is integrating globalization and AI deeply into its new strategic focus. Nie Lei stated that in the first half of this year, overseas revenue accounted for over 40%, with projections indicating it will exceed 50% by 2026. In global competition, Changfei’s benchmark has shifted from Japanese peers to U.S.-based Corning.
China CITIC Telecom Mobile is targeting a position among the leading suppliers of 6G technology. To achieve this, the company plans to invest 10 billion RMB in R&D over the next five years, maintaining R&D expenditure at 15% to 20% of revenue.
The Optics Valley is also striving to create a more comprehensive support system for industrial innovation. Yu Song, Director of the Donghu High-Tech Zone Management Committee, emphasized that unwavering support for enterprises is the lifeline for navigating economic cycles. The region will continue refining patient capital mechanisms, talent policies, and industrial infrastructure to foster thriving enterprise clusters.
Source: AI Observer
Disclaimer: Contains third-party opinions, does not constitute financial advice
HuggingFace's "AI Duck" Goes Viral, Boosting Chinese Chip Supplier Rockchip
09-02
14 Executives Depart, Horizon Unleashes a Robot Legion
08-11
Nokia Snatches AI Optical Chip Race, Stock Soars 52%
08-07
AI stocks down 40%-60%, yet NVIDIA's valuation is at a 10-year low
08-06
HBM Bottleneck Migration, NVIDIA Also Aims to Downgrade
08-06
AI Servers Fuel Demand for Minor Metals: Prices Surge Over 70% in Half a Year
08-06
AI is not a search bar—it's a supercomputer costing just a few dozen dollars
08-05






