The regulatory backdrop
U.S. regulators are reportedly working on measures to bar imports of new Chinese optical transceivers into the American market, with officials aiming to publish the rules sometime in 2026. Banking and sector research from BofA has emphasized that the proposals are still fluid and could be significantly revised or even withdrawn prior to finalization. Two unresolved definitional issues stand out and will determine the practical scope of any rule set.
Unresolved definitional questions
First is whether enforcement would hinge on the manufacturing location of the devices or on the country of domicile of the company that supplies them. Under a location-based approach, products made outside China could avoid restrictions even if supplied by Chinese-owned vendors. Under a domicile-based approach, products from Chinese firms would be targeted regardless of where they were manufactured.
Second is whether the restrictions, if enacted, would reach backward to cover transceivers already being shipped today - for example, 800G and 1.6T modules - or whether they would be limited to future product generations. Each interpretation carries markedly different consequences for operators, vendors, and the timeline for any supply disruptions.
Until the FCC clarifies these two issues, any precise calculation of impact on supply chains and customer deployments remains premature.
How concentrated is the market?
The optical transceiver market is heavily dominated by non-U.S. suppliers. More than 70% of global supply originates from vendors based outside the United States, with the overwhelming majority of that volume produced by Chinese manufacturers. The supplier ecosystem is layered, featuring distinct tiers of vendors.
- Tier-1: Innolight, Eoptolink
- Tier-2: CIG, DS Precision (Source Photonics), Accelink, HG Tech
- U.S. vendors with China exposure: Coherent (COHR), which uses Chinese subcontractors Hui Lyu and USI for overflow manufacturing capacity
Manufacturing footprints are geographically dispersed across Thailand, Malaysia, China, and Taiwan for both Chinese-headquartered and U.S.-headquartered vendors. That distribution matters because it complicates any enforcement mechanism that is tied solely to production location.
Southeast Asia expansion and its implications
Chinese suppliers have already started to shift capacity outside mainland China. Leading names in the sector began expanding operations into Southeast Asia in 2024 with the express purpose of serving U.S. demand. According to channel checks cited by BofA, by 2025 the non-China facilities of leading Chinese vendors were capable of meeting U.S. demand on a standalone basis. This geographic hedging could blunt the effectiveness of a restriction linked to manufacturing location and would complicate enforcement tied to factories rather than corporate ownership.
Next-generation datacenter development and strategic risk
A critical and often underappreciated factor is the co-development pipeline for next-generation datacenter optics. U.S. cloud service providers continue to partner with Chinese vendors on multiple fronts, including next-generation pluggable optics (NPO), 2.4T and 3.2T transceiver designs, and other emerging datacenter interconnect architectures. A broad import ban or limits on collaborative product development would not only perturb current component flows, but could also delay deployment of those next-generation architectures at a time when AI infrastructure buildout is accelerating.
Technology and national security framing
Optical transceivers perform signal conversion; they do not process or store user data. That technical distinction narrows the national security rationale for restrictions compared with semiconductors or general networking equipment. Nevertheless, the market is concentrated, and there is limited near-term ability for U.S. firms to substitute at scale. Domestic companies face tight supply conditions and lack comparable production capacity at the scale required today.
What investors and operators should watch
Regulators could still change direction, and the final regulatory language will determine whether any restriction is tied to the place of manufacture or the vendor's domicile and whether it reaches products already in shipment. Given the current state of uncertainty, the regulation is best characterized as an evolving tail risk for participants in AI infrastructure, datacenter networking, and the optical components sector. It does not represent an immediate, crystallized supply shock until the FCC releases specific, actionable language.
Close monitoring of rule text is essential. If the FCC specifies company-domicile-based controls or includes products already being shipped, the implications for procurement and deployment timelines could be far more acute than under a narrow, future-products-only, location-based approach.