The Chip Is Now
the Border

On October 13, FCC Report and Order 26-50 took effect. From that morning, a new radio-frequency device cannot get U.S. equipment authorization if any logic-bearing component inside it was produced by Huawei, ZTE, or another Covered List entity. The name on the box no longer decides the question. The chip does.

What happened

The order, published in the Federal Register on September 11, 2026, closes what Chairman Brendan Carr called the component part loophole. Before October 13, the FCC could block a finished device made by a Covered List company, while a different maker could source a baseband processor, Wi-Fi chipset, or IoT module from the same company and still pass. The new definition sits at 47 C.F.R. Section 2.902: a component that generates timing signals above 9,000 pulses per second, or uses radio-frequency energy to process data. Passive parts are excluded. Programmable parts are in.

Carr has put the principle in four words: trust starts in silicon. Garmin and Sony asked the Commission for a bright line list of banned part numbers. The FCC refused and kept a totality of circumstances test instead. Who designed the logic, who holds the IP, and who controls fabrication all count. Buying through a distributor in between changes nothing.

Why it matters

Pending applications are grandfathered as filed. Amend one to swap a logic-bearing chip and the filing counts as new, with the full bar attached and no exception for a forced substitution during a shortage. Certification bodies such as UL, TUV Rheinland, Nemko, and SGS now have to run a provenance review before they grant authorization. That turns a radio interference check into an import gate, and it puts the cost on the applicant who cannot document a clean chain.

My position: own the suppliers who can prove provenance on paper, and be careful with the importers who assumed provenance was someone else's problem. Qualcomm (QCOM) sells the cellular and connectivity silicon that device makers will reach for when a Covered List part has to come out of a design. Every redesign forced by this rule is a sales call QCOM did not have to make. The honest caveat is timing. Existing authorizations stay valid, so the revenue arrives on redesign cycles, not overnight.

The one number

9,000 pulses per second. That threshold, borrowed from the FCC definition of a digital device at 47 C.F.R. Section 15.3(k), is now the line between a component the rule ignores and a component that can keep a product out of the United States. It is a low line, set on purpose, and it captures most of the computing layer in a modern device.

What to watch

The companion rulemaking closed its comment record on September 28, 2026. Still open in it: mandatory hardware bills of materials with country of origin for every logic-bearing component, a ban on Covered List software and firmware, and a wider ban on all components from those entities. The Semiconductor Industry Association backed the security goal and opposed itemized disclosure, proposing attestation under NIST SP 800-161r1 instead. Watch how the FCC resolves that fight, and watch the marketplace deadlines that follow: March 1, 2027 for sellers holding inventory, June 1, 2027 for third-party platforms. If the bill of materials proposal advances, October 13 will read in hindsight as the mild version.