FCC Chinese Optical Module Rules: What Changed Oct 13

In August, we covered the reported FCC effort to restrict Chinese optical transceivers in US data centers, and the disposition question almost no one was asking: what happens to the enormous installed base of these components when it retires? Since then the picture has clarified in an important and widely misunderstood way. The FCC’s component rule went live on October 13, 2026, but the Chinese optical module makers everyone expected to be hit were not added to the Covered List, and the next restriction being discussed targets a future generation of hardware, not what is in the racks today. Here is what actually changed, what did not, and why the distinction matters for anyone who will eventually retire this hardware.

Note: This article is educational and describes an evolving regulatory environment with a finalized general rule, an unlisted set of specific vendors, and a separate still-forming proposal. It is not legal advice, and ROC Telecom is not a law firm. Anyone making procurement or disposition decisions involving potentially restricted components should consult qualified export-control and regulatory counsel. Details reflect reporting and FCC records as of October 19, 2026.

TL;DR

The regulatory picture has three layers, and keeping them straight is the whole point. As of mid-October 2026:

The component rule is now live. The FCC’s Third Report and Order (FCC 26-50) took effect October 13, 2026, closing the “component-part loophole” by allowing the agency to deny equipment authorization to devices that contain logic-bearing hardware components produced by a Covered List entity.
Chinese optical module makers were not added to the Covered List. Zhongji Innolight, Eoptolink, Dongshan Precision, and TFC were all left off, and the order notes that major Chinese optical module makers are not on the list. Because the rule keys off whether a component’s maker is on the Covered List, it does not currently restrict Chinese optical transceivers.
The next restriction being reported targets the 3.2T generation, not today’s hardware. According to reporting tied to Morgan Stanley meetings in Washington, any transceiver-specific FCC action would phase in starting with the 3.2T generation, which is expected to ramp around 2028 to 2029, explicitly leaving current 800G and 1.6T deployments out of the initial focus.
A 65 percent US-content carve-out is under discussion. The reported framework would let Chinese-manufactured modules through if at least 65 percent of their bill-of-materials value comes from US companies, a threshold a US-sourced DSP paired with US-sourced lasers could approach.
A separate Senate bill is moving in parallel. A bipartisan bill reported in late September would bar Chinese optical transceivers (naming Innolight and Eoptolink) from federal national-security systems, with a five-year transition, a narrower scope than the data center market as a whole.

The disposition takeaway, sharper than before: the current installed base of Chinese optical transceivers in US data centers is now doubly insulated from the regulatory action, because the makers are not listed and the forward-looking restriction aims at a generation that is not yet deployed. That hardware will keep running and retire on its normal refresh cycle into the secondary market, exactly as the installed-base dynamic we described in August anticipated. The regulatory trajectory is now clear enough to plan disposition around rather than guess at.

What Went Live on October 13

The most concrete development is that the FCC’s component rule is no longer a proposal. It is in effect.

On July 22, 2026, the FCC adopted its Third Report and Order (FCC 26-50, ET Docket 21-232), and after Federal Register publication, the rule took effect on October 13, 2026. What it does is close what the FCC called the “component-part loophole.” Previously, the Covered List prohibitions applied to finished equipment: a complete device from a listed entity could be denied authorization, but a device from a non-listed maker that merely contained a listed entity’s component could pass. From October 13, that distinction closes. The FCC can now deny new equipment authorization to a device if it incorporates a logic-bearing hardware component produced by a Covered List entity, where the device would itself have been prohibited had that entity produced the whole unit.

Several limits are built into the rule, and they matter for reading it correctly. It applies to new authorizations and does not revoke existing ones. Applications pending on the effective date are exempt unless later amended to change a logic-bearing component. It does not cover software or firmware. And a listing based only on production location does not, by itself, trigger the component prohibition. The rule is a mechanism, in other words, a tool the FCC can now use, rather than a specific ban on any particular product.

This is the piece of the story that is genuinely settled. The mechanism we described in August as the distinctive feature of this effort, the FCC equipment-authorization and Covered List regime rather than the Commerce Department export controls that govern GPUs, is now operative law.

What Did Not Change: The Part Most Coverage Missed

Here is where the nuance lives, and where a lot of the alarmed August coverage turned out to overshoot. The component rule going live did not, in fact, restrict Chinese optical transceivers, because of who is and is not on the Covered List.

The Major Chinese Module Makers Stayed Off the List

The component rule only bites when a device contains a logic-bearing component from an entity that is actually on the Covered List. And the major Chinese optical module manufacturers were not added. Zhongji Innolight, Eoptolink, Dongshan Precision, and TFC Optical Communication were all left off, and a footnote in the order confirms that major Chinese optical module makers are not on the Covered List. Optical modules appear in the rulemaking only as an example of a logic-bearing component, not as a separately designated restricted item, and the companies that dominate the market were not named.

The practical effect is that, as of mid-October 2026, the live component rule does not restrict the import or authorization of Chinese optical transceivers. The mechanism exists; the specific vendors it would need to target to hit optical modules are not on the list that activates it. The supply chain, which had braced for a blanket restriction, got a reprieve.

The Installed Base Is Untouched, Again

This reinforces the central point from our August analysis. Nothing about the October 13 rule, or the vendors left off the list, pulls a single Chinese transceiver out of a US data center. The enormous installed base of Chinese optical modules already deployed across American AI infrastructure keeps running exactly as before. The rule is forward-looking and authorization-based; it governs what can be newly authorized, not what is already installed and operating. The millions of Chinese transceivers in the data path of US facilities remain in place, and will remain in place until they retire on the normal refresh cycle.

What Is Next: The 3.2T Generation, Not Today’s Hardware

The most forward-looking development, and the one that most sharpens the disposition picture, is the shape of the restriction now reportedly being discussed. It is aimed at a future generation of hardware.

The 3.2T Focus

According to reporting tied to Morgan Stanley meetings in Washington, any FCC transceiver-specific restriction would phase in starting with the 3.2T generation of optical modules, rather than a blanket ban on all Chinese optics at once. Crucially, the current 800G and 1.6T products are reported as not the focus. AI data centers are in the middle of large 800G and 1.6T deployments right now, and a 3.2T-first approach is described as insulating today’s volume drivers while shaping where the next generation gets built. The 3.2T generation itself is expected to ramp around 2028, with broader adoption around 2029, which means the restriction, if it lands, would govern a product cycle that is still roughly two years out.

The 65 Percent US-Content Rule

The reported framework includes a notable carve-out. Chinese-manufactured transceivers could still be imported if at least 65 percent of their bill-of-materials value comes from US companies. Morgan Stanley’s reported assessment is that a US-sourced digital signal processor paired with US-sourced lasers would bring a module close to that threshold. The threshold itself, the legal definition of a “US company,” and the schedule are all described as still under deliberation and not yet codified. This is a proposal taking shape, not a rule, and the details that would determine its real effect are unsettled.

The Parallel Senate Bill

Separately, a bipartisan group of senators proposed a bill in late September that would bar Chinese optical transceivers, naming Innolight and Eoptolink, from federal national-security systems, with a five-year transition and authority for Commerce and Defense to expand the list. This is a narrower scope than the commercial data center market: it targets federal national-security systems specifically, not every US data center. It is worth tracking as a separate track from the FCC action, with its own, more limited, reach.

How the Market Read It

The equity markets have moved on each turn of this story, which is a useful signal of where supply-chain advantage is seen to be shifting. Western and Taiwanese optical suppliers have risen on restriction headlines through 2026, with US names like Applied Optoelectronics, Lumentum, and Coherent repeatedly cited as potential beneficiaries, and Taiwanese laser and component suppliers rallying on the 3.2T reporting. Lumentum, for its part, has told the FCC that any policy should be phased in and warned that an abrupt cutoff would cause supply-demand dislocation. The market reaction underscores that this is understood as a gradual, forward-looking redistribution of the optical supply chain, not an overnight ban.

Why This Clarifies the Disposition Picture

For anyone who will eventually retire this hardware, the net of these three developments is not more uncertainty. It is more clarity. The regulatory trajectory is now defined enough to plan around.

Current-Generation Chinese Optics Are Doubly Insulated

The installed base of Chinese optical transceivers in US data centers, overwhelmingly 800G and 1.6T, is now insulated from the regulatory action on two counts. First, the makers are not on the Covered List, so the live component rule does not touch them. Second, the forward-looking restriction is aimed at the 3.2T generation that is not yet deployed, explicitly leaving current-generation products out of its initial focus. The hardware operators are running today is not the target of either the live rule or the reported proposal. It keeps operating, and it will retire on the normal refresh cycle into the secondary market.

The Retirement Stream Is Predictable, Not Frozen

In August, the open question was whether a hardening posture would complicate the disposition of already-installed Chinese optics. The October picture answers much of that. The current-generation installed base is not restricted, so its retirement into resale, redeployment, and recycling proceeds under normal conditions, not under a cloud of new prohibition. The regulatory pressure is being directed at the next generation and at new builds, not at the retirement stream of what is already deployed. For disposition planning, that means the large wave of Chinese optical modules that will retire over the coming refresh cycles can be handled as the valuable secondary-market hardware it is, rather than as suddenly-stranded assets.

Provenance Discipline Still Pays

None of this removes the value of knowing exactly what is installed. The regulatory environment is still sharpening its focus on component origin and vendor identity, the Senate bill names specific Chinese makers, and the 3.2T framework turns on bill-of-materials content and US-company definitions. An operator who maintains serialized inventory and can document hardware by vendor, type, and generation is positioned to navigate whatever the 3.2T rule becomes, to answer procurement and compliance questions with confidence, and to route each category of retiring hardware to its appropriate disposition path. The clarity in the current picture is a reason to get the provenance discipline in place now, while the forward-looking rules are still forming, not a reason to relax it.

What Operators Can Reasonably Do Now

The three-layer picture suggests a measured posture: act on what is settled, track what is forming, and keep the asset-management discipline that serves either way.

Treat the current installed base as normal secondary-market hardware. The live rule does not restrict current-generation Chinese optics, and the reported proposal targets a future generation. Retiring 800G and 1.6T Chinese transceivers can proceed as ordinary high-value disposition, through resale, redeployment, and certified recycling.
Know what is installed, by vendor and generation. Maintain serialized inventory that identifies optical modules by maker, type, and speed generation. This is the foundation for answering any future question the 3.2T rule or the Senate bill raises, and for routing retiring hardware correctly.
Track the 3.2T proposal on its own timeline. The reported restriction concerns a generation ramping around 2028. There is time to watch how the 65 percent content threshold, the “US company” definition, and the schedule resolve before they affect procurement or disposition decisions. Monitor, do not overreact.
Distinguish the federal-systems track from the commercial market. The Senate bill targets federal national-security systems, a narrower scope than general data center operations. Operators serving federal workloads should track it specifically; others can weigh it as context rather than an immediate constraint.
Keep export-control screening in the disposition plan. For any cross-border movement of retired optical hardware, the export-control considerations covered elsewhere in our reporting still apply, independent of the FCC’s domestic authorization rules. Involve qualified counsel where cross-border movement or significant procurement is involved.

The unifying idea has not changed since August: good disposition and asset-management discipline is exactly the capability a sharpening component-security environment rewards. What has changed is that the environment is now defined enough that operators can plan with confidence rather than brace for an unknown.

Frequently Asked Questions

The following is general educational information about an evolving regulatory environment, not legal advice.

Did the FCC ban Chinese optical transceivers on October 13, 2026?

No. What took effect on October 13 was the FCC’s component rule (its Third Report and Order, FCC 26-50), which closes the “component-part loophole” by letting the agency deny new equipment authorization to devices containing logic-bearing hardware components produced by a Covered List entity. But the major Chinese optical module makers, including Innolight, Eoptolink, Dongshan Precision, and TFC, were not added to the Covered List. Because the rule only applies when a component’s maker is on that list, it does not currently restrict Chinese optical transceivers. The mechanism is live, but the specific vendors it would need to target to hit optical modules are not listed.

Were Innolight and Eoptolink added to the FCC Covered List?

No. As of mid-October 2026, Zhongji Innolight, Eoptolink, Dongshan Precision, and TFC Optical Communication were all left off the Covered List, and the FCC’s order notes that major Chinese optical module makers are not on it. They are, however, named in a separate bipartisan Senate bill that would bar Chinese optical transceivers from federal national-security systems, which is a distinct track from the FCC Covered List and narrower in scope, covering federal national-security systems rather than the commercial data center market.

What is the 3.2T optical module restriction?

It is a restriction reportedly being discussed, not yet a rule, that would phase in starting with the 3.2T generation of optical transceivers rather than banning all Chinese optics at once. According to reporting tied to Morgan Stanley meetings in Washington, current 800G and 1.6T products are not the initial focus, and the 3.2T generation is expected to ramp around 2028 to 2029. The reported framework includes a carve-out allowing Chinese-manufactured modules if at least 65 percent of their bill-of-materials value comes from US companies. The threshold, the definition of a “US company,” and the schedule are all described as still under deliberation and not codified.

Does any of this affect Chinese optical modules already installed in US data centers?

No. Neither the live October 13 component rule nor the reported 3.2T proposal pulls existing hardware out of service. The component rule governs new equipment authorization, not deployed equipment, and the Chinese module makers are not on the Covered List in any case. The reported 3.2T restriction targets a future generation that is not yet deployed, explicitly leaving current 800G and 1.6T products out of its initial focus. The large installed base of Chinese optical transceivers in US data centers keeps operating and will retire on the normal refresh cycle, unaffected by either action.

What does this mean for retiring Chinese optical transceivers?

It means the current-generation installed base can be retired as ordinary high-value secondary-market hardware, not as stranded or restricted assets. Because the makers are unlisted and the forward-looking restriction targets a future generation, the 800G and 1.6T Chinese optics now in US data centers are not subject to the regulatory action, and their retirement into resale, redeployment, and certified recycling proceeds under normal conditions. The practical discipline that matters is serialized inventory and provenance documentation, so hardware can be identified by vendor and generation and routed to the right disposition path, which positions an operator for whatever the 3.2T rule eventually becomes.

Is the component rule the same as the GPU export controls?

No. They are different legal mechanisms. The FCC component rule works through the agency’s equipment-authorization and Covered List regime and governs what can be imported into and authorized for use in the United States. The GPU export controls are administered by the Commerce Department’s Bureau of Industry and Security under the Export Administration Regulations (the ECCN 3A090 framework) and govern what advanced hardware can be exported out of the United States. The export controls restrict what leaves the country; the FCC rule restricts what can be authorized to come in and be sold. Both can apply to the same hardware at different stages of its life, especially at cross-border disposition.

How does the 65 percent US-content rule work?

As reported, the proposed framework would allow Chinese-manufactured optical modules to be imported if at least 65 percent of their bill-of-materials value comes from US companies. Morgan Stanley’s reported assessment is that a module using a US-sourced digital signal processor and US-sourced lasers would come close to that threshold, which puts the practical pressure on the optical laser and DSP supply chain. Importantly, the 65 percent threshold, the legal definition of a “US company,” and the implementation schedule are all described as still under deliberation and not yet codified, so the real-world effect remains to be determined.

Should operators stop buying Chinese optical modules now?

That is a procurement decision for each operator and its counsel, but the current facts do not compel an abrupt change. The live rule does not restrict Chinese optical modules, the makers are not on the Covered List, and the reported future restriction targets the 3.2T generation ramping around 2028 with a content-based carve-out still being defined. The measured posture reported across the market is phased adaptation rather than an overnight cutoff, which is also what Western suppliers like Lumentum have urged on the FCC. Operators should track how the 3.2T framework resolves and maintain provenance discipline, while recognizing that current-generation procurement is not currently restricted.

How is this different from what was reported in August?

In August, the reporting described a possible blanket FCC ban on new Chinese optical transceivers, and the framing was uncertain and alarmed. The October picture is more defined and more measured. The general component mechanism is now final and live as of October 13, but the Chinese module makers were not listed, so optical transceivers are not currently restricted, and the transceiver-specific action being discussed is a phased, 3.2T-generation-first proposal with a US-content carve-out, aimed at future hardware rather than today’s installed base. The trajectory is clearer, narrower in immediate effect, and more forward-looking than the August coverage suggested.

The Bottom Line

The story we covered in August has resolved into a clearer and more measured shape than the early coverage implied, and the distinctions matter. The FCC’s component rule went live on October 13, 2026, giving the agency a real tool to deny authorization to devices containing logic-bearing components from Covered List entities. But the major Chinese optical module makers were not added to that list, so the live rule does not currently restrict Chinese optical transceivers at all. And the transceiver-specific restriction now being reported is a phased proposal aimed at the 3.2T generation ramping around 2028, with a 65 percent US-content carve-out still under deliberation, explicitly leaving today’s 800G and 1.6T deployments out of its initial focus. A separate Senate bill targets federal national-security systems specifically, a narrower track again. Three layers, three different scopes, and only one of them is actually in effect.

For infrastructure operators and anyone who will retire this hardware, the practical message is clarity rather than alarm. The current installed base of Chinese optical transceivers in US data centers is insulated from the regulatory action on two counts, the makers are unlisted and the forward-looking restriction targets a future generation, so it keeps running and retires on the normal refresh cycle into the secondary market, exactly as the installed-base dynamic we described in August anticipated. The regulatory pressure is being directed forward, at the next generation and at new builds, not at the retirement stream of what is already deployed. The operators who maintain serialized inventory and provenance discipline now are positioned to handle that retirement wave with confidence, and to navigate whatever the 3.2T rule becomes when its details are finally codified.

How ROC Telecom Helps

ROC Telecom is an R2v3, RIOS, NIST 800-88, and ITAR-compliant ITAD specialist built for exactly the visibility and disposition discipline this layered regulatory environment rewards:

Serialized, per-asset inventory and chain-of-custody that identifies and documents optical modules and other hardware by vendor, type, and generation, the foundation for navigating component-origin questions
Specialist asset recovery across optical transport, transceivers, routing, switching, and compute, with direct buyer relationships and a deep domestic US secondary-market network for current-generation hardware retiring on the refresh cycle
Export-control-aware disposition for hardware whose cross-border movement raises screening questions, integrated with denied-party and destination review
NIST 800-88 data destruction with serialized Certificates of Destruction and full chain-of-custody documentation
R2v3 Appendix E materials recovery with in-house dismantling and direct-to-refiner processing for genuine end-of-life components
Nationwide decommissioning with 48-hour rapid-response mobilization

15+ years of ITAD experience, $25M+ in client capital recovered, 45M+ pounds diverted from landfill.

ROC Telecom does not provide legal advice and takes no position on pending regulation. We help operators retire and recover data center hardware with the documentation and disposition discipline an evolving security environment demands.

Request a Free Infrastructure Disposition Assessment

Tell us about the optical transport and data center hardware you are retiring or evaluating. A specialist will discuss recovery value, secure data destruction, provenance documentation, and disposition strategy suited to a sharpening component-security environment. No commitment, no spam.

Prefer to talk directly? Call 585-406-1249 or email info@roctelecom.com.

"*" indicates required fields

Related reading:

What Changed on October 13, and What Didn’t: The FCC’s Chinese Optical Module Rules