The FCC has proposed opening unlicensed bands to satellites, and asked what should protect the incumbents

The proposal puts satellites in the bands Wi-Fi and industrial IoT already share

The US Federal Communications Commission has proposed allowing unlicensed devices to communicate directly with satellites in three bands: 902-928 MHz, 2400-2483.5 MHz and 5725-5850 MHz ^[1][2]. The item was adopted on 6 August 2026 as FCC 26-51 in ET Docket No. 26-169 and published in the Federal Register on 8 September 2026 at 91 FR 57110 ^[3][10]. Comments are due on or before 9 November 2026 and reply comments on or before 7 December 2026 ^[3].

Those three bands are where Wi-Fi, Bluetooth, industrial telemetry and most unlicensed IoT already operate ^[1][7]. The Commission selected them precisely because they permit relatively higher unlicensed power ^[7][10], and together they amount to more than 200 megahertz of spectrum ^[8]. The proposal would let equipment certified under Part 15 transmit to satellites without a separate satellite licence, under a licensing-by-rule approach modelled on the Supplemental Coverage from Space framework adopted in 2024 ^[1].

The operative condition is that satellite operations in these bands would be unprotected. They “may not cause harmful interference to authorized services and must accept any interference received” ^[1]. That is the standard Part 15 bargain, and it has held for decades. It has held because every emitter in the bargain was terrestrial and short-range.

What the current rules assume

The technical baseline is 47 CFR 15.247. Digitally modulated systems are capped at 1 watt maximum peak conducted output power in all three bands, with antennas assumed not to exceed 6 dBi directional gain, and higher-gain antennas requiring a corresponding power reduction ^[4]. Out-of-band emissions in any 100 kHz must sit at least 20 dB below the highest in-band level in 100 kHz ^[4]. Power spectral density for digitally modulated systems is capped at 8 dBm in any 3 kHz band ^[4].

Every one of those numbers encodes an assumption about geometry. A 1 W transmitter with a 6 dBi antenna interferes with things nearby. Coexistence in unlicensed spectrum works because interference is local, so the population of devices that can affect each other is bounded by distance. A transmitter in orbit does not have that property. One beam illuminates a footprint measured in tens or hundreds of kilometres.

The Commission has asked the interference questions rather than answered them

This is the part practitioners should read carefully. The FCC states an intent to “preserve the operational status quo of the spectral environment for incumbent and authorized services” ^[2]. It has not proposed the limits that would achieve it. It has asked for them.

On aggregate effects the Commission asks: “Is there a need for an aggregate field strength or PFD limit to maintain the continued operational utility of the part 15 ecosystem and protect authorized users?” ^[2] It separately seeks comment on “whether there should be any limit on the number of beams allowed to illuminate the same geographical area simultaneously” ^[2]. On downlinks it asks whether power flux density limits similar to those in Part 25 would be appropriate and sufficient to protect incumbent federal radiolocation operations ^[2]. On emissions it asks what out-of-band limits would be necessary for space-to-Earth operation, and whether existing Part 15 limits suffice or Part 25 standards should apply ^[1][2].

It also asks the blunt version directly: would allowing Part 15 devices to transmit Earth-to-space “increase the likelihood of harmful interference to authorized services or adversely affect spectrum sharing by unlicensed devices?” ^[2] The Commission references a European precedent of -142 dB(W/m2) in a 4 kHz bandwidth while leaving open whether such a limit suits US conditions ^[1]. Coexistence mechanisms floated for comment include geographic restrictions to underserved areas, beam footprint limits, aggregate emission caps across constellations, and listen-before-talk ^[1].

Europe went a different way, and the contrast is instructive

CEPT’s ECC Report 373, approved on 13 February 2026, divides direct-to-device into two classes ^[5]. D2D-MSS uses established Mobile-Satellite Service allocations in L and S band, and the report concludes that “the existing MSS framework is ready to accommodate compliant D2D-MSS without regulatory changes” ^[5]. D2D-IMT, which reuses terrestrial mobile bands from roughly 698 to 2690 MHz, does not have that luxury: it currently depends on derogations under Radio Regulations No. 4.4, and the report records that “no existing limits and coordination triggers in RR Articles 21 and 22, have been developed for protection of incumbent satellite and terrestrial services operating in specific MFCN frequency bands” ^[5].

Europe’s route therefore runs through licensed spectrum, operator partnerships and coordination procedures ^[5][6]. The FCC proposal is a third path: unlicensed, unprotected, and without the coordination machinery that licensed sharing provides. Neither approach is settled. The ITU is considering allocations under WRC-27 agenda items 1.12, 1.13 and 1.14 ^[6], and the conference will examine direct satellite links to standard handsets across spectrum from 694/698 MHz to 2.7 GHz ^[9].

Why this matters outside the United States

An FCC rule binds the United States. The equipment market does not respect that boundary. Devices certified for one regime are manufactured once and sold across regions, and a chipset built to transmit to satellites in 2.4 GHz will be the same chipset shipped elsewhere ^[10]. More importantly, the technical record built in this docket will feed the WRC-27 positions that ITU member administrations, including those across APAC, then have to answer ^[6][9]. The argument made in these comments is the argument regulators will be handed in 2027.

What to do

• Note the dates. Comments close 9 November 2026 and replies 7 December 2026 in ET Docket No. 26-169 ^[3]. An organisation with a material dependency on these bands has a narrow window to put evidence on the record.
• Inventory what you actually run in 902-928 MHz, 2.4 GHz and 5725-5850 MHz. Industrial telemetry, asset tracking, building systems and process instrumentation often sit in these bands and are rarely in an RF inventory at all ^[4][7].
• Treat the aggregate PFD question as the one that decides the outcome. A per-satellite limit and an aggregate limit across constellations are different protections, and the Commission has asked about both without proposing either ^[2].
• Record a noise-floor baseline in your bands now, while the environment is unchanged. Without a before, no organisation will be able to evidence an after.
• Read Part 15 correctly before relying on it. Unlicensed operation carries no protection from interference and an obligation to accept it ^[1][4]. That is the existing position, not a change.
• Track WRC-27 agenda items 1.12, 1.13 and 1.14 alongside the FCC docket. The international decision, not the national one, is what reaches operators in this region ^[6][9].

Domain close

Unlicensed spectrum has been one of the more successful pieces of radio regulation, and it worked on an unstated assumption: that every participant in the bargain was close enough to the others for interference to be a local, bounded problem. A satellite is not close to anything and is close to everything. The FCC has recognised this precisely enough to ask whether aggregate power flux density limits and beam-count caps are needed, which is the right question at the right stage. Nothing here is in force, and nothing here is settled. But the electromagnetic environment that industrial wireless quietly depends on is now the subject of an open proceeding, and the record closes in December.