Goldman Sachs: Three Key Debates on China's Optical Transceiver Ban — Industry Leaders' Dominance Hard to Shake
Taylor Wilson
The US is drafting a ban on Chinese-made optical modules for data centers. Goldman Sachs laid out three core debates around the potential impact, concluding that AI demand and accelerating tech cycles are deepening customers' reliance on incumbent leaders — making the industry hierarchy hard to disrupt near-term.
Is this ban real, or a bargaining chip?
The Trump administration is drafting legislation to ban Chinese data-center component imports; the FCC is studying parallel measures targeting optical modules.
B. Riley analysts speculate the ban may never land — it could be a diplomatic lever ahead of a Trump–Xi summit.
This means → the ban carries a "bid high, then negotiate" flavor. Markets need to separate the policy signal from the final outcome.
Why does faster tech iteration make customers *more* dependent on incumbents?
Optical modules span multiple speeds (1.6T, 2.4T, 3.2T), form factors (pluggable, LPO, CPO), and material platforms (silicon photonics, EML, lithium niobate) — the SKU count is enormous, and R&D barriers are steep.
In plain terms = an optical module is not one standard product but a collection of dozens of custom variants. The faster technology moves, the harder it is to switch suppliers — switching costs rise with complexity.
Goldman notes that global leaders were first to ship 400G (2018), 800G (2020), and 1.6T (2023) transceivers. That "first to volume, first to revenue" track record keeps extending.
How strong is China's edge in capacity and cost efficiency?
Seven of the world's top-ten optical-module vendors by revenue are headquartered in China, and their combined share grew further from 2024 to 2025.
These leaders design their own production lines and proprietary equipment, keeping manufacturing know-how and automation expertise in-house — difficult for outsiders to replicate.
This means → upgrading to 1.6T and beyond requires packing more fibers and lasers into the same module footprint, sharply raising coupling and thermal demands. Smaller suppliers need more time to catch up, so the leaders' advantage is actually widening.
How fast can the supply chain move out of China?
Goldman expects leading vendors to expand manufacturing in Thailand and elsewhere in Southeast Asia to hedge geopolitical risk.
This fits the broader industry trend of regionalized supply chains — not exiting China, but building a second leg outside it.
This reflects a bet by the leaders themselves: whether the ban lands is a variable, but pre-positioning diversified capacity is the hedge against that variable.
What is Goldman's bottom line on these leaders?
Although the bank's trading desk takes no position on the ban's outcome, Goldman maintains a positive view on its covered names — FOCI, RoboTechnik, Landmark, Eoptolink, and VPEC.
The core logic is three lines converging: strong AI demand + tight raw-material supply + accelerating tech cycles — all three deepen customers' reliance on top-tier suppliers.
Put simply = the ban is a policy risk, but the leaders' moat is built on technology and capacity. Policy can reroute trade, but it cannot quickly change who can build — and who can build fast.
Content is for reference only, not financial advice.