645 Thought Leadership

Reading the FCC's Chip Rule: A Prediction and an Investment Map

In July 2026, the FCC released its Third Report and Order on protecting the communications supply chain through the equipment authorization program. Most of the coverage treated it as a ban on components from five Chinese companies. That reading is accurate, and it misses the point. Lined up against the last five years of FCC action, this order is the fourth step in a clear sequence, and hardware founders in the US should be planning for the fifth.

What the rule actually does

The order bars devices containing "logic-bearing hardware components" tied to entities on the Covered List, the FCC's roster of companies and equipment deemed a national security risk. The definition is broad: any device, module, sub-assembly, integrated circuit, or other physical component that "generates and uses timing signals or pulses at a rate in excess of 9,000 pulses per second and uses digital techniques." That sweeps in semiconductors, IoT modules, optical transceivers, and populated circuit boards. Excluded are structural parts, passive components, basic power-only parts like a plain battery cell, and bare motors with no embedded digital controller.

Read the definition twice and it becomes clear what it is. Nine thousand pulses per second plus digital techniques is a description of a chip. The FCC has written the definition of a chip into federal rules and reserved the right to decide whose chips get in.

Five Covered List entities are directly implicated: Huawei, ZTE, Hytera, Hikvision, and Dahua. All five sell chips, modules, and processors that can end up inside a third party's product. Until now, a US device maker could buy a module from one of them, solder it onto a board, and sell the finished product under its own name. Starting October 13, that path closes for new devices seeking FCC authorization. For Huawei and ZTE parts, that means essentially any product sold in the US. For Hikvision, Dahua, and Hytera, it means devices used for public safety, government facilities, critical infrastructure, and national security.

The pattern matters more than the rule

The Covered List began in 2021 as a list of names. If Huawei made it, it was covered. In 2022 the FCC stopped authorizing new equipment from listed companies altogether. Sharper teeth, same logic.

2025 broke that logic. The Commission expanded the Covered List to prohibit all UAS, UAS critical components, and routers produced in a foreign country, regardless of who made them. For the first time, covered equipment was defined by production location rather than by provider. The rule came with waivers through January 2027 for equipment on the Blue UAS Cleared List and for anything qualifying as a "domestic end product." The waivers drew most of the attention, but the shift in principle was the real news.

The July order takes the next step and reaches inside the device to the component level. In the same document, the FCC seeks comment on splitting the Covered List into producer-based and production-location-based categories, and on applying the rules sector by sector, effectively creating "Covered List sectors." Taken together, those two questions describe a Commission preparing to stop naming companies and start naming categories.

My prediction: the FCC moves to ban all foreign-produced logic-bearing components, with exceptions and waivers for the first 18 to 24 months, for devices used for public safety, government facilities, critical infrastructure, and national security. I see the drone critical components rule as the FCC's template: draw the line by geography, then manage the transition with carve-outs that expire. I predict a wide range of waivers and exceptions in the chips space.

Why this could outlast any one administration

The obvious objection is that this is one administration's project and could be reversed. My answer is maybe. The statute underpinning the Covered List was signed in 2020, the list itself was published in 2021, the Secure Equipment Act passed Congress with near-unanimous support, and the 2022 authorization ban and the 2025 location-based expansion came from FCCs led by opposite parties. There is bipartisan support around the importance of secure communications and devices, but different beliefs in how to effectively enforce this. I think how this ban goes into effect, and the impact on US industry will be telling.

The regulator's incentives could point the same way. Designating companies one at a time is slow and easy to route around with a subsidiary or a white-label arrangement. Every new incident pushes the Commission toward a rule it can administer without chasing corporate structures, and production location is that rule. The first signal to watch is January 2027, when the drone waivers come up.

Where I'm looking as an investor

The useful exercise is mapping what the rule touches and what it leaves alone, because that is where the pressure, and the opportunity, will land.

Circuit boards: assembly is exposed, fabrication is not. A bare PCB, even an advanced HDI or substrate-grade board, is copper traces, laminate, and plated vias. No clock, no digital logic, no data processing until something is soldered onto it. My prediction is that this next iteration of the FCC rule will reach the logic-bearing components, the assembly step, and the finished populated board. It will not reach design, design software, or bare-board fabrication. If the location-based logic extends to boards, the openings are in US assembly (a notoriously thin-margin business), compliance and traceability tooling, quality assurance, and component sourcing.

Chips: skip the fab, win the layer after it. New US fab capacity is needed, but it is the hardest, slowest, most capital-intensive layer to win, and TSMC and Samsung have committed to building it here alongside Intel. The layer this rule quietly makes more valuable is US-based, traceable packaging and assembly. How "origin" gets defined for a chip is still unsettled: US customs has traditionally looked to where a chip is packaged, and China now looks to where the wafer was made. Either way, packaging is where provenance gets recorded and proven. Provenance becomes a feature of the part. Advanced packaging also has a second tailwind. As transistors get harder to shrink, performance gains increasingly come from stacking, chiplets, and heterogeneous integration. The incumbents own high-volume packaging, and they're the ones building it here. The startup openings sit around the factory rather than in it: new substrate materials, die-to-die interconnects, and the tools to prove every die in a package works (including metrology and inspection).

Optical transceivers: named outright. A transceiver generates and processes digital signals, so it sits squarely inside the definition. I'm watching three layers. Light generation is the bottleneck, and the interesting work is in new laser architectures: III-V-on-silicon integration, quantum dot lasers, and comb lasers that generate many wavelengths from one source instead of one laser per channel. Photonic integration packs more optical functions onto a single chip, so transceivers get cheaper and denser without a matching jump in power. Co-packaged and near-packaged optics move the optics closer to the switch silicon to cut the electrical losses that mount at higher data rates. As compliance costs rise for anything with foreign logic inside it, US-designed and US-traceable versions of each become more valuable.

The July FCC order isn't the end point. If the FCC follows its own trajectory, the question for every US hardware company will shift from who made its chips to where they were made. I'm looking for the founders building for that question now.