If you’re a global institutional investor, venture capitalist, or private equity professional looking at China’s optical transceiver industry, you’re not alone. The numbers are extraordinary, the growth trajectory is steep, and the strategic importance of this sector is increasingly recognized by investors worldwide.

But here’s the challenge: China’s optical module market is complex, fast-moving, and full of nuance. Market leadership doesn’t always translate to investment viability. Technology leadership doesn’t always mean commercial success. And geopolitical risks can shift the landscape overnight.

This guide provides a comprehensive overview of the investment landscape in China’s optical transceiver industry — the market dynamics, the investment themes, the key players, and the risks you need to consider before deploying capital.

---

The Investment Thesis — Why Now?

AI computing power is the single largest driver of the optical transceiver industry, and the industry is firmly in a growth phase with no signs of slowing[reference:0]. According to LightCounting, the global Ethernet optical transceiver and CPO market reached $16.5 billion in 2025 and is projected to hit $26 billion in 2026 — representing a staggering 60% growth rate for two consecutive years[reference:1]. The optical module market is expected to reach approximately $38 billion in 2026, with growth exceeding 35% year-over-year[reference:2].

Looking further ahead, the global optical module market is projected to exceed $76 billion by 2030, representing more than tenfold growth over the decade[reference:3]. The compound annual growth rate from 2026 to 2030 is estimated at around 20%, with AI-dedicated optical modules growing even faster[reference:4].

Chinese manufacturers now dominate the global supply chain, accounting for approximately two-thirds of global optical transceiver supply and about 60% of global revenue. Chinese firms claim seven spots among the world’s top ten optical module vendors[reference:5]. In the first four months of 2026, Chinese companies’ global market share crossed the 50% threshold for the first time, reaching 52%[reference:6].

The concentration trend is clear: the top five players command a 61.4% market share, with Chinese firms dominating the upper echelons. Industry analysts expect market concentration to increase further as R&D and expansion capital requirements rise, creating a pronounced “Matthew effect” where the strong get stronger[reference:7].

---

Four Investment Themes — Where to Focus

1. Speed Upgrade — 800G → 1.6T → 3.2T

Product iteration cycles have compressed from 3–4 years to approximately 2 years. This accelerated upgrade cycle creates sustained demand for next-generation products and pricing power for first movers.

800G has become the industry’s core growth driver, with shipments expected to more than double in 2026[reference:9]. 1.6T is experiencing explosive commercial takeoff in 2026, with shipments growing from a small base in 2025 to tens of millions of ports[reference:10]. 1.6T chipset sales are expected to exceed $2 billion in 2026[reference:11]. 3.2T is the next frontier, with LightCounting expecting gradual ramp-up from 2028 onward[reference:12].

According to LightCounting, the combined 800G and 1.6T optical module market is expected to reach $14.6 billion in 2026, accounting for approximately 64% of the total Ethernet optical module market[reference:13]. 1.6T optical module shipments are projected at 15 million units in 2026, with a market size of $4.5 billion[reference:14]. Citigroup has further raised its forecast to 22 million units[reference:15].

2. Localization — Upstream Import Substitution

The upstream supply chain remains heavily dependent on overseas sources, creating significant import substitution opportunities. In high-speed optical modules, optical chips can account for 50% to 70% of total costs, and high-end EML chips and Indium Phosphide substrates remain largely imported[reference:16].

Current supply bottlenecks include Indium Phosphide substrates, EML chips, DSP chips, Faraday rotators, and silicon lenses — all critical components with constrained capacity[reference:17]. The global EML supply-demand gap has widened to over 30%, with orders booked through 2028[reference:18].

Chinese companies are actively investing in upstream localization. 源杰科技 (Yuanjie Technology) is a leader in high-speed EML optical chips[reference:19]. 光迅科技 (Accelink) is the only domestic company with InP optical chip IDM capabilities, integrating upstream chips with downstream modules[reference:20]. 光库科技 (Gkoptics) specializes in thin-film lithium niobate (TFLN) modulators — a next-generation technology for ultra-high-speed applications[reference:21].

3. Platformization — Vertical Integration as a Moat

Vertical integration is becoming a critical competitive advantage. AI customers have increasingly high requirements for delivery speed and quality, and supply chain stability is paramount. Companies that control their own optical chip design, packaging, coupling, and testing — improving yield and delivery capability while reducing costs — have a significant edge.

中际旭创 (InnoLight) exemplifies this trend. The company’s construction-in-progress surged from 156 million yuan to 2.36 billion yuan — a 14-fold increase — reflecting aggressive capacity expansion across the value chain[reference:22]. 光迅科技 operates a complete self-developed production line covering optical chips, optical devices, and optical modules[reference:23].

4. Ecosystem Building — NPO, CPO, and the Next Architecture

New technology ecosystems are taking shape around NPO (Near-Package Optics), CPO (Co-Packaged Optics), LPO (Linear Pluggable Optics), and silicon photonics[reference:24].

2026 marks the first year that silicon photonics-based transceivers will exceed 50% of total market revenue[reference:25]. The deployment of NPO and CPO is accelerating silicon photonics adoption and driving demand for InP continuous-wave lasers[reference:26].

Google has placed a formal order for 12 million NPO optical modules for its next-generation TPU v7/v8 clusters, with Chinese suppliers capturing nearly all of it[reference:27]. NVIDIA has announced that its Spectrum-X Ethernet silicon photonic switch — the world’s first mass-produced 200G/lane CPO Ethernet switch system — has entered full production, with laser count reduced by 75%, power consumption cut by 80%, and MTBF improved 10x[reference:28].

---

Targets Worth Watching — The Investment Landscape

Tier 1 — Global Leaders (AI Computing Core)

中际旭创 (InnoLight, 300308.SZ) — The world’s No. 1 optical transceiver supplier, with over 40% share in the 800G market and an estimated 50% to 70% share in the 1.6T segment[reference:29]. The company is a core supplier to NVIDIA, Google, and Meta, with silicon photonics and EML dual technology paths[reference:30]. In Q1 2026, the company reported single-quarter net profit of 5.7 billion yuan, far ahead of peers[reference:31]. Order coverage extends through all of 2026, with some orders already placed for 2027[reference:32]. The company’s gross margin surged from 36.70% to 46.06% year-over-year[reference:33].

新易盛 (Eoptolink, 300502.SZ) — The world’s No. 2 supplier, specializing in LPO low-power solutions[reference:34]. The company has secured Amazon and Meta as key customers and has established overseas production capacity to hedge against tariff risks[reference:35]. Gross margins consistently exceed 48%[reference:36]. With a market capitalization roughly half that of InnoLight, Eoptolink offers lower volatility and comparable growth prospects[reference:37]. 1.6T production capacity is expected to accelerate in the second half of 2026[reference:38].

Tier 2 — Upstream Champions and Niche Leaders

光迅科技 (Accelink, 002281.SZ) — The only domestic InP optical chip IDM, with integrated upstream chip and downstream module capabilities[reference:41]. Strong in telecom long-haul and coherent modules[reference:42]. 1.6T modules are now capable of mass delivery[reference:43]. However, the company has lower share among North American AI hyperscalers, making it less elastic than Tier 1 players[reference:44]. Valuation caution: TTM PE of 212x is the highest in the sector[reference:45].

源杰科技 (Yuanjie Technology, 688498.SH) — The leading supplier of high-speed EML optical chips, the “heart” of optical modules[reference:46]. The company is a critical upstream player in the supply chain[reference:47]. High-speed EML chips are a key bottleneck in the industry, creating significant import substitution opportunities[reference:48].

光库科技 (Gkoptics) — A specialist in thin-film lithium niobate (TFLN) modulators, one of the most promising technologies for beyond-1.6T applications[reference:49]. TFLN offers superior bandwidth, lower power consumption, and better signal integrity than traditional materials[reference:50].

“Little Giants” — Hidden Champions in the Supply Chain

天孚通信 (TFC, 300394.SZ) — A leader in optical engines and passive components, supplying nearly all major module manufacturers[reference:51]. The company has strong financial performance with excellent cash flow[reference:52]. Valuation caution: TTM PE of 169x is significantly high, with growth rates lower than module manufacturers[reference:53].

联特科技 (Linktel, 301205.SZ) — A high-speed optical module specialist with full portfolio coverage[reference:54]. The company is a founding member of the XPO MSA and debuted a 12.8T XPO module at OFC 2026[reference:55].

鑫巨宏 (Xinju Hong) — A precision optical device supplier, holding 64 patents (including 17 invention patents) and certified as a national-level “Little Giant” enterprise[reference:56]. The company’s IPO application on the Beijing Stock Exchange was accepted in December 2025[reference:57]. It is a critical supplier to InnoLight and other global leaders[reference:58].

恩达通 (Endatong) — A national-level “Little Giant” enterprise with nearly 100 patent applications[reference:59].嘉元科技 (Jiayuan Technology) invested 500 million yuan for a 13.59% stake in the company in 2026[reference:60].

---

Investment Risks — What to Watch

1. Technology Iteration Risk

The industry is evolving at breakneck speed. Technology roadmaps can shift suddenly, and companies that bet on the wrong technology path can be left behind[reference:61]. LPO, NPO, CPO, silicon photonics, TFLN, and traditional EML are all competing for dominance[reference:62]. CPO technology could potentially replace pluggable optical modules in the long term, resetting the competitive landscape[reference:63].

2. Supply Chain Bottlenecks

Upstream supply constraints are the single largest near-term risk to the industry’s growth trajectory[reference:64]. Indium Phosphide substrates, EML chips, DSP chips, Faraday rotators, and silicon lenses are all in critical shortage[reference:65]. InP substrate demand is projected to reach 3 million wafers in 2026, but effective capacity remains severely constrained[reference:66]. These bottlenecks could limit production even as demand surges, creating a “capacity race hitting a supply ceiling” scenario[reference:67].

3. Policy and Geopolitical Uncertainty

The U.S. Federal Communications Commission is drafting measures to restrict imports of new Chinese-made optical transceivers, with officials hoping to finalize and implement the rules within 2026[reference:68]. While the proposal faces significant industry resistance and is still in draft form, geopolitical risk is real and growing[reference:69]. Policy implementation timelines could be delayed, modified, or abandoned — but investors need to monitor developments closely[reference:70].

4. Valuation Risk

The optical module sector is trading at historically high valuation levels. During the 2023 market peak, leading players traded at only 50–60x PE. Now, TTM PEs across the sector are generally above 70x, reflecting significant sentiment premiums for 1.6T and 3.2T iteration expectations[reference:71]. Key downside triggers include: hyperscaler capex falling short of expectations, sharp price declines in 1.6T modules compressing gross margins, CPO technology replacing pluggable modules (lowering the long-term ceiling), and a pullback in U.S. AI tech stocks dragging down Chinese growth stocks[reference:72].

---

The Bottom Line — Due Diligence Is Non-Negotiable

China’s optical transceiver industry presents one of the most compelling growth stories in the global technology sector. AI-driven demand, accelerating product cycles, upstream localization, and new technology ecosystems are creating significant investment opportunities across the value chain.

But the same factors that create opportunity — speed, complexity, and concentration — also create significant due diligence challenges. Market leadership doesn’t guarantee investment viability. Technology leadership doesn’t guarantee commercial success. And geopolitical risks can shift the landscape overnight.

Before you invest, you need to know:

• Is the company’s reported market share accurate? — or is it marketing hype?
• Does the company actually own its core technology? — or is it licensed from others?
• What is the company’s real financial health? — beyond the headline numbers?
• Who are the company’s key customers and suppliers? — and how dependent are they?
• What are the company’s legal and compliance risks? — any disputes, violations, or regulatory issues?

These are the questions that publicly available data alone cannot answer. And they’re the questions that professional due diligence exists to address.

ChinaBizInsight — Know Your Chinese Partners. Visit our website to learn more about our full range of China business intelligence services.