Q: Why is this year’s optical communication investment strategy clearly shifting upstream?
A: The core change this year is the increasingly obvious supply-demand mismatch for upstream materials, with profits shifting to the segments facing greater shortages.
Demand for optical modules is growing rapidly—some segments may see tenfold growth—but expanding production of upstream materials like optical chips, Faraday rotators, and Indium Phosphide substrates is difficult. This mismatch will intensify competition and lead to price increases.
Therefore, investment strategy must move further upstream. Optical chips, core passive components, companies with optical chip capabilities, and second-tier module makers that can secure critical supplies all deserve greater attention.
Q: What is the logic behind optical chip price increases?
A: Severe supply-demand imbalance, especially for EML and CW lasers.
A 70mW CW laser has risen from ~$3.50 to $3.80. A 200G EML has gone from ~$9–10 to $15, and could reach $20. Lasers are not the highest-cost component, but they are critically scarce—shortages affect shipments, creating strong pricing power. As 1.6T, silicon photonics, and CPO/NPO demand grows, EMLs, CW lasers, and high-power lasers will become even tighter bottlenecks.
Q: How are Chinese optical chip companies categorized?
A: Two directions: EML and CW lasers.
EML: Changguang Huaxin, Source Photonics (mainly self-supplies, ~11 MOCVD reactors), Chenguang (200G EML possible by late 2025/early 2026).
CW lasers: Innolight (Yuanjie Technology) has the strongest capabilities—large-scale delivery, positive validation with Innolight, smooth overseas expansion. High-power lasers are its next growth driver. Shijia Photonics and Yongding Co. are also worth watching.
Q: Why does Indium Phosphide substrate affect global optical chip supply?
A: Going further upstream from optical chips leads to Indium Phosphide substrates, then phosphorus and indium. If substrates are constrained, EML and CW production capacity suffers.
China accounts for a large share of global output for these materials, with relatively strict export restrictions. Wafer fabs in Taiwan and Japan could be affected. If China tightens controls further, the global supply-demand mismatch for optical chips will widen, putting upward pressure on prices—a long-term positive for Chinese optical chip manufacturers.
Q: Why are Faraday rotators becoming a bottleneck?
A: Their value share in a module is not high, but they are a critical component—without them, shipments stop.
The global Faraday rotator market is dominated by Coherent (US) and a Japanese firm, which together control ~90%. Chinese manufacturers exist, but capacity is limited. As demand ramps up, leading module makers will book supply early. Second-tier manufacturers face shipment risks if they cannot secure stable supply. The core issue is certainty of supply, not unit price.
Q: What is the most difficult part of optical module manufacturing?
A: Optical path coupling. It relies heavily on skilled technicians working under a microscope with a six-axis motion platform and an optical power meter—adjusting in real time until the reading peaks, then applying adhesive to fix alignment.
This process is highly dependent on manual experience, making efficiency, yield, and consistency challenging. Many module makers have historically relied on large pools of skilled workers and process expertise.
Q: Why are automation capabilities critical for the future of optical communication?
A: The future is not just 800G → 1.6T → 3.2T linear upgrades. It demands higher precision, greater integration, and more complex packaging—large-scale silicon photonics, CPO/NPO, higher-order optical interconnects, and semiconductor-like packaging and testing.
As the industry evolves, demands for high-precision pick-and-place, coupling, bonding, packaging, and automation will only increase. Automation is not just for 1.6T mass production today; it enables the next several generations of optical interconnect technology.
Q: Why are global leaders entering the optical communication market?
A: Because the market space, profits, and long-term trend are too attractive to ignore.
Recent moves: Dongshan Precision (via Source Photonics), TSMC (via packaging capabilities), Google, Meta, Nvidia, Broadcom—all are entering optical interconnects with their own strengths.
Optical communication is no longer a small cycle within traditional telecom. It is core infrastructure for AI’s long-term expansion. As long as the future potential of optical interconnects is ten times today’s size, there is room for major module makers, companies with optical chip capabilities, and new entrants with automation and packaging expertise—all can have their place.
