• Microsoft (NASDAQ:MSFT) is lobbying against a proposed U.S. ban on Chinese optical transceivers, warning of supply-chain disruptions that could stall AI data-center construction.
  • The company has 35 U.S. data centers under construction and another 134 planned, while China controls roughly two-thirds of the global optical transceiver market.
  • A possible compromise is under consideration: restricting next-generation 3.2-terabit-per-second devices rather than current technology.

Microsoft Pushes Back on Proposed Ban

Microsoft is urging the Trump administration to abandon plans to ban Chinese optical transceivers, arguing that cutting off essential networking components could severely delay U.S. AI infrastructure projects. The software giant has 35 data centers under construction and another 134 planned, according to project tracker Cleanview, and relies on Chinese suppliers who dominate the global market for the fiber-optic components that connect servers and switches.

The warning comes as officials weigh security concerns against the rapid buildout of AI capacity. China accounts for roughly 67% of the global optical transceiver market, according to Counterpoint Research, and U.S. competitors like Coherent (NYSE:COHR) and Lumentum (NASDAQ:LITE) lack the production scale to replace Chinese suppliers, according to a Foundation for American Innovation report cited by Reuters (NASDAQ:TRI).

A Possible Compromise

The most important new detail is a potential compromise: officials are reportedly considering targeting next-generation 3.2-terabit-per-second devices rather than imposing an immediately disruptive restriction on current technology. Production of those devices is not expected to ramp until 2028, with widespread installation expected in 2029. Two sources said that threshold is under consideration, though the FCC has not publicly confirmed it.

Microsoft and other companies met with the White House Office of Science and Technology Policy and National Security Council in September to warn about supply-chain disruption, according to Reuters. Microsoft declined to comment publicly on the report.

Financial Stakes

The financial exposure is significant. Microsoft reported $92.3 billion in quarterly revenue for the period ended September 30, with $39.4 billion in cash additions to property and equipment, up from $19.4 billion a year earlier. Commercial remaining performance obligations—contracted revenue not yet recognized—rose 73% to $677 billion. Delays in securing components could postpone the conversion of that spending and those commitments into revenue.

ITI general counsel John Miller warned that already-purchased AI chips could sit unused without the networking equipment needed to connect them. His warning represents the industry lobby’s assessment, not an independently established forecast of a nationwide shutdown.

The market has already treated restrictions as potentially favorable to domestic suppliers: after Reuters’ August 4 report, Lumentum, Coherent and Applied Optoelectronics (NASDAQ:AAOI) shares rose approximately 7%, 11% and 18%, respectively.

Political and Industry Context

Two separate policy initiatives should not be conflated. The FCC initiative would restrict imports of new Chinese optical-transceiver models, with a next-generation speed threshold under consideration. A bipartisan congressional bill introduced September 25 targets Chinese transceivers in sensitive federal government systems, initially naming Eoptolink (SZSE:300502) and Zhongji Innolight (SZSE:300308), and would allow five years for compliance if approved.

Officials and lawmakers cite potential espionage, malware and service-disruption risks. Reuters also notes Zhongji Innolight’s inclusion on a Pentagon list of alleged Chinese military-backed companies. These are security concerns and designations; the reviewed reporting does not establish that the transceivers at issue have actually been used to compromise U.S. data centers.

The international-policy tension is straightforward: Washington wants less dependence on Chinese technology while accelerating U.S. AI development, but an abrupt restriction could undermine the infrastructure expansion supporting that goal. Microsoft and industry groups argue that slower deployment would instead benefit China.

What’s Next

The most consequential next development is the FCC’s actual text: which suppliers and device generations it covers, when restrictions begin, and what exemptions or transition arrangements it allows. Until those details are public, “could halt expansion” is best understood as a serious industry warning rather than a settled forecast.

An expansive restriction covering technology needed today would present the greatest disruption risk. A narrower rule could substantially change that outcome. The reported 3.2 Tbps approach could give alternative suppliers time to expand before next-generation deployments become widespread, but the central uncertainty is whether that capacity can arrive fast enough to meet demand without materially delaying AI investment.

Update: This article was updated to clarify that Microsoft’s data-center figures come from project tracker Cleanview and that the 3.2 Tbps threshold is under consideration but not confirmed by the FCC.