Goldman Sachs: US Import Restrictions Unlikely to Shake Leading Chinese Optical Module Makers

Overview

This report addresses recent media reports suggesting the US government may prohibit imports of Chinese data center components (including optical modules). It outlines three core debates of greatest concern to investors. Goldman Sachs believes that despite macroeconomic uncertainty, cloud service providers (CSPs) rely even more heavily on existing Chinese industry leaders against the backdrop of surging AI demand, accelerated technology iteration, and raw material shortages. The report details why new entrants struggle to replace leaders in the short term, the sources of China's cost and efficiency advantages in manufacturing, and how Southeast Asian capacity layouts mitigate tariff risks, ultimately maintaining Buy ratings on relevant targets.

Core views

Technology Iteration and Customer Stickiness Form Core Barriers: The research report argues that in the current environment of rapidly evolving AI server architectures, CSPs find it difficult to diversify their supply chains. Optical modules are undergoing rapid migration from 800G to 1.6T, 3.2T, and new forms such as LPO/CPO, involving various materials (SiPh, EML, etc.) and fiber types, resulting in extremely complex SKUs. This high technical threshold requires suppliers to possess strong R&D capabilities and early involvement. Chinese leading enterprises, leveraging long-term cooperation with CSPs, can participate in product design early, whereas new suppliers need longer periods to verify product stability and integrate into the ecosystem. Data shows that seven of the global top ten optical module manufacturers by revenue are located in China, further expanding their market share in 2025, confirming deepening industry reliance on existing leaders rather than weakening. Manufacturing Efficiency and Scale Advantages Are Hard to Replicate: Beyond R&D strength, Chinese leading enterprises have significant advantages in capacity commitment, automated production, and manufacturing efficiency. As optical module speeds increase to 1.6T and above, integrating more lasers and fibers without increasing volume imposes extremely high requirements on coupling precision and heat dissipation, increasing manufacturing difficulty. Leading enterprises keep automated production know-how internal through self-developed production lines and proprietary equipment, thereby achieving competitive prices and fast delivery while ensuring yield rates. The report notes that during this window of rapid technological evolution, small manufacturers are unlikely to reach comparable manufacturing efficiency levels in a short time. Geographic Diversification of Supply Chains Has Substantially Materialized: Facing ongoing geopolitical tensions since 2019, the pandemic, and increased tariffs amid macroeconomic uncertainties, the Chinese optical module supply chain has proactively pursued geographic diversification. Leading enterprises have established bases capable of producing high-end products like 1.6T in Southeast Asian regions such as Thailand and are expanding capacity in other countries to further disperse risks. For example, Eoptolink's first phase of capacity in Thailand is fully operational, with the second phase being expanded in 2026. This forward-looking overseas layout allows leading enterprises to effectively buffer potential impacts from US import restrictions and ensure continuous supply to global customers.

Analysis framework

The research report adopts an analytical framework combining 'event-driven' and 'industrial fundamentals'. First, it uses the unexpected event of potential US import restrictions as an entry point to identify the expectation gap between market sentiment and industrial reality. Second, it delves into the micro-level of the industry chain, quantifying the actual feasibility of 'domestic substitution' or 'supply chain de-Sinicization' by analyzing the pace of optical module technology iteration (speed, packaging, materials), SKU complexity, and CSP validation cycles. Third, combining manufacturing attributes, it breaks down cost moats from perspectives of automated equipment, yield ramp-up, and economies of scale. Finally, by tracking the progress of specific companies' overseas factory construction (e.g., Thailand factory utilization rates), it verifies the effectiveness of risk hedging measures, concluding that fundamentals remain robust despite negative sentiment.

Methodology notes

• Competitive & Strategic FrameworkMoat / competitive advantage

  Technology Iteration Speed as a Switching Cost Barrier

  In industries with rapid technological upgrades (such as AI optical modules), existing leaders accumulate substantial tacit knowledge and validation data by participating in clients' early R&D; new entrants not only need to catch up on technical parameters but also pay high time costs to prove reliability. This high switching cost driven by 'speed' constitutes a dynamic moat.

• Industry / Industrial Analysis FrameworkUpstream-Midstream-Downstream Transmission

  Reverse Lock-in of Upstream Supply Chain Structure by Downstream CSP Demand

  It is generally believed that trade restrictions weaken the export country's share, but during the special period of AI computing power shortage, downstream CSPs prioritize delivery certainty and technical matching over geopolitical compliance costs, causing demand-side reinforcement of binding with mature head suppliers. This reflects the reverse shaping power of urgent downstream demand on upstream supply structure.

• Company Fundamentals & Financial FrameworkOperating / Financial Leverage Analysis

  Operating Leverage Driven by Specialized Automated Equipment

  Optical module leaders solidify core processes into equipment through self-developed proprietary production equipment. Although initial capital expenditure is high, marginal costs decrease significantly as output expands; simultaneously, this specialized equipment itself serves as a physical barrier to prevent technology leakage and maintain yield rate advantages.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Eoptolink (易飞扬)

  Beneficiary: Leader in AI optical modules, leading overseas capacity layout

  Strengths

  Thailand factory Phase 1 fully operational, Phase 2 expanding in 2026; leading progress in 1.6T product R&D and mass production

  Comparison

  Faster implementation speed of high-end capacity in Southeast Asia compared to peers, stronger risk resistance

  Risks

  Overseas expansion progress falling behind expectations; unexpectedly stringent US tariff policies

• FOCI (富创光电)

  Beneficiary: Leader in optical communication components, deeply bound to CSP clients

  Strengths

  Long-term service to global CSPs, early involvement in new product R&D; comprehensive SKU coverage

  Comparison

  Traditional advantages in customized services and customer stickiness

  Risks

  Technology route change risks; tight raw material supply

• RoboTechnik (罗博特科)

  Beneficiary: Provider of optical module automated equipment and manufacturing solutions

  Strengths

  Provides proprietary automated equipment, helping leaders build manufacturing efficiency barriers

  Comparison

  Scarcity in the field of optical module-specific automation

  Risks

  Fluctuations in downstream capital expenditure; equipment obsolescence due to technology iteration

• Landmark (联钧光电)

  Beneficiary: Optical transceiver module manufacturer, benefiting from increased industry concentration

  Strengths

  Strong R&D capabilities, product line covers high-speed modules

  Comparison

  As one of the industry leaders, enjoys economies of scale advantages

  Risks

  Intensified market competition; exchange rate fluctuations

• VPEC (全新光电)

  Beneficiary: Optical chip and module supplier, vertical integration advantages

  Strengths

  Accumulation in specific materials and chips, benefiting from silicon photonics penetration

  Comparison

  Differentiated competitiveness in upstream core component segments

  Risks

  Technology path dependency risks; customer concentration risks

Key data

• Global Optical Module Market Size Forecast2026E: $50.9 Billion; 2027E: $72.6 BillionDriven by AI demand, global TAM is expected to grow significantly in 2026-2027, followed by a slight correction to $69.1 Billion in 2028E
• High-Speed Optical Module Shipment Share2028E: 1.6T accounts for 34%, 3.2T accounts for 31%By 2028, products at 1.6T and above will become the main shipment force, accounting for over 65% combined, indicating extremely rapid technology iteration
• Silicon Photonics (SiPh) Penetration Rate2025: 28%; 2028E: 62%The penetration rate of silicon photonics solutions will rise rapidly from 28% in 2025 to 62% in 2028, becoming the mainstream technology route
• Chinese Manufacturers' Market Position7 out of Global Top 10By revenue, seven of the global top ten optical module suppliers are Chinese enterprises, and their share continued to expand in 2025 compared to 2024
• Eoptolink Thailand Capacity ProgressPhase 1 full production, Phase 2 expansion in 2026As a typical case of supply chain diversification, its Thailand factory Phase 1 is operating at full load, with Phase 2 capacity under construction

Impact & implications

For the optical module industry, the report's judgment implies that even if the US implements import restrictions, Chinese leading enterprises are likely to maintain their core position in the global AI supply chain due to technological generational gaps and overseas capacity layouts. For investors, this indicates that current geopolitical panic may present mispricing opportunities, as the fundamental support logic (AI demand, technological barriers, manufacturing efficiency) remains unbroken. Simultaneously, this implies increased differentiation within the industry: only leaders possessing global manufacturing capabilities and next-generation technology R&D strength can navigate the cycle, while smaller and medium-sized enterprises lacking these capabilities will face greater survival pressure.

Risks

• Actual US import restriction measures are stricter or broader in scope than expected
• Slowing growth in AI capital expenditure leads to optical module demand falling short of expectations
• Yield ramp-up or capacity release at Southeast Asian overseas factories slower than planned
• Route changes in next-generation technologies (such as CPO/NPO) leading to replacement of existing products
• Continuous shortages of key raw materials (such as laser chips) affecting deliveries

What to watch

• Specific details and timelines regarding US FCC and Commerce Department rules on Chinese optical module import restrictions
• Certification progress and order guidance from major CSP clients for 1.6T/3.2T optical modules
• Factory capacity utilization rates and expansion announcements for Chinese leaders in Thailand, Vietnam, etc.
• Global AI server shipment volumes and chip platform iteration pace
• Changes in the actual penetration rate of Silicon Photonics (SiPh) technology among leading manufacturers