The Trump administration's push to cut China out of the U.S. AI supply chain might have an unintended consequence: making it harder and pricier to build the very AI infrastructure American companies are scrambling to deploy.
That's the gist of a new note from BNP Paribas, which argues that a proposed Federal Communications Commission ban on new imports of Chinese-made optical transceivers could send shockwaves far beyond Chinese manufacturers, hitting companies like Broadcom Inc. (AVGO), Marvell Technology Inc. (MRVL), Credo Technology Group Holding Ltd. (CRDO), Nvidia Corp. (NVDA), and Advanced Micro Devices Inc. (AMD).
Why Optical Transceivers Matter to AI
Optical transceivers don't usually make headlines, but they're a critical piece of modern AI infrastructure. These devices convert electrical signals into optical signals, letting the massive amounts of data generated by AI chips zip between servers inside data centers.
As AI clusters grow, these networking components become just as important as the GPUs doing the heavy lifting. Without enough optical transceivers, even the most advanced AI chips can't communicate efficiently across thousands of interconnected servers.
According to BNP Paribas analyst Karl Ackerman, Chinese manufacturers are expected to control more than 60% of the global data center optical transceiver market in 2026, with Innolight and Eoptolink among the biggest suppliers serving the U.S. market.
The Ban Could Reach Beyond Chinese Companies
At first glance, restricting Chinese suppliers seems like a golden opportunity for non-Chinese competitors. But Ackerman argues the supply chain is far more tangled than that simple story suggests.
Optical transceivers rely on digital signal processors, or DSPs—specialized chips that handle the high-speed data traveling through fiber-optic networks. BNP estimates those processors account for roughly 40% of a transceiver's bill of materials.
That means Chinese transceiver makers are also major customers of DSP suppliers like Broadcom, Marvell, Credo, and MaxLinear Inc. (MXL). If Chinese production gets curtailed, those chipmakers could lose a significant chunk of demand.
The ripple effects could spread even further. Laser suppliers such as Lumentum Holdings Inc. (LITE) and Coherent Corp (COHR) also sell components into Chinese transceiver manufacturing, and companies supplying semiconductor materials and manufacturing equipment could see weaker demand if production slows.
Nvidia and AMD Could Face Another AI Constraint
The implications go beyond component suppliers.
Ackerman notes that optical transceivers from Innolight and Eoptolink are widely used in AI server systems built around Nvidia's upcoming Vera Rubin platform and AMD's MI450 accelerators.
Replacing that manufacturing capacity won't be easy. BNP argues U.S. suppliers are unlikely to immediately produce enough 800G and emerging 1.6-terabit transceivers to meet demand as hyperscalers ramp up next-generation AI infrastructure.
Instead of strengthening the domestic AI ecosystem, the bank believes an outright ban could tighten supplies of critical networking equipment, raise infrastructure costs, and add more inflationary pressure to already record AI capital spending.
The timing is especially crucial. BNP forecasts the data center optical transceiver market will grow 41% year over year to roughly $31 billion in 2026, driven by the rollout of Nvidia's Vera Rubin systems, AMD's MI450 platform, Alphabet Inc. (GOOGL)'s TPU8 processors, and Amazon.com Inc. (AMZN)'s Trainium 3 chips.
For investors, the report highlights an increasingly important reality of the AI buildout: the next supply-chain bottleneck may not come from GPUs, but from the networking hardware that connects them.
