Compliance 📅 September 6, 2026 ⏱ 11 min read

The FCC 900 MHz Proceeding: What to Specify in a UHF RFID Reader for the US in 2026

A UHF RFID reader on a test bench beside a spectrum display showing the 902 to 928 MHz band divided into hop channels

<p>Most retail chains running RAIN RFID leave FCC dockets to their suppliers. In 2026 this one is worth reading directly. A petition to re-cut the 902–928 MHz band that every US UHF RFID reader depends on has moved far enough through the Commission that a draft rulemaking notice went into interagency review this spring, and a coalition of 60 companies and trade associations filed against it in June. The band stands as it is today, and it may well stay that way. The distance between “as it stands today” and “order adopted” is exactly the window in which multi-year purchase orders get signed. Here is what the proceeding proposes, what has been filed against it, and the specific clauses worth adding to a US reader specification while it runs.</p>

What NextNav asked the FCC for, in plain terms

NextNav filed its petition for rulemaking with the FCC in April 2024. It asks the Commission to restructure the lower 900 MHz band — the 26 MHz from 902 to 928 MHz that RAIN RFID shares with smart meters, alarm radios, tolling transponders, LoRa, Wi-SUN and a long tail of Part 15 devices — in order to host a terrestrial 5G-based positioning, navigation and timing service, positioned as a complement and backup to GPS. The petition runs as RM-11989 and WT 24-240; the Commission’s separate proceeding on PNT technologies is WT 25-110.

The geometry is the part worth memorising, because everything else follows from it:

A 5 MHz uplink at 902–907 MHz — the bottom edge of the band.
A 10 MHz downlink at 918–928 MHz — the top edge.

That is 15 of the band’s 26 MHz, taken from both ends at once. Simple arithmetic leaves 907–918 MHz — 11 MHz — as the remaining contiguous middle, and the allocation of that middle is one of the open questions the proceeding would have to settle.

Take the edges and you change the band, not just its size. A US reader hops across the whole band. Under 47 CFR 15.247, a frequency-hopping system whose 20 dB channel bandwidth is under 250 kHz uses at least 50 hopping frequencies and holds an average occupancy of no more than 0.4 seconds on any one frequency in a 20-second period. That hop table is spread deliberately across the full 26 MHz, and dense-reader installations — a dock wall, a store ceiling grid, a conveyor line — depend on that spread to keep neighbouring readers clear of each other. Compress the usable span and you compress the separation available to every reader in the building at once. Both the hop plan and the front-end filter that shapes it are designed around today’s band edges.

Power is the second half of the picture. RFID Journal reports that the proposed power levels reach an EIRP density of 3,280 W per MHz — a different order of magnitude from the Part 15 devices sharing the same spectrum, where the ceiling is 1 W maximum peak conducted output, rising to roughly 4 W EIRP with a directional antenna. Industry filings in the docket argue that readers within about a kilometre of such a transmitter could be desensitised. Those assertions sit in the record of a live proceeding, still to be weighed — which is precisely why the technical record has become the centre of the fight.

Who filed against it, and what they actually filed

Three filings matter to anyone specifying RFID hardware, and they do different jobs.

The coalition letter — RILA, 3 June 2026

The Retail Industry Leaders Association led 60 businesses and trade associations urging the FCC to reject the reconfiguration. The signatory list is unusually broad for a spectrum docket, and its breadth is the argument: the RAIN Alliance, GS1 US and AIM Global from the identification industry; the National Retail Federation and FMI from retail; the Consumer Technology Association, the Information Technology Industry Council, TechNet and the Telecommunications Industry Association from tech; the Wi-Fi Alliance and the Dynamic Spectrum Alliance from wireless; the Security Industry Association, the Electronic Security Association and The Monitoring Association from alarms; Edison Electric Institute, the American Gas Association and the American Petroleum Institute from utilities and energy; the American Trucking Associations plus more than a dozen state trucking associations, E-ZPass and the International Bridge Tunnel and Turnpike Association from transport; several aviation bodies; and the US Chamber of Commerce.

Read that list as an inventory of what actually sits in the band. RFID is one tenant among many, and the coalition’s framing is that the band quietly carries infrastructure that runs out of sight.

The identification industry response — 15 May 2026

The RAIN Alliance, GS1 US and AIM Global issued a joint response to NextNav’s coexistence testing, concluding that the materials presented “fall far short of addressing the documented coexistence and interference risks” in the lower 900 MHz band. Their objection is symmetrical, and that symmetry is the interesting part: they raise the risk to RAIN RFID from high-power terrestrial operations, and equally the reverse case — a very large installed base of Part 15 devices operating into a planned 5G receiver estate. Mutual interference is a harder engineering problem than one-way interference, and the filings ask for both paths to be modelled.

The technical assessment — the Plum Consulting report

Filed by a consortium including the RAIN Alliance alongside the LoRa Alliance, Wi-Fi Alliance, Wi-SUN Alliance and Z-Wave Alliance, this is a methodology challenge rather than a policy plea. What the consortium asks the Commission to require of the coexistence analysis:

5G base station density modelled at realistic deployment levels — the filing argues the petition’s figure sits up to an order of magnitude below them.
Part 15 devices modelled as the listen-before-talk radios they are, backing off rather than transmitting continuously.
Outdoor-specific analysis to support the assumed 70/30 indoor/outdoor split.
Adjacent-channel interference: the consortium asks for it to be modelled.
Reciprocal interference from Part 15 devices into 5G receivers: requested as a required part of the analysis.
Outdoor infrastructure — smart meters, agricultural sensors, street lighting: the consortium asks for a dedicated assessment.

Procedurally this is a different kind of intervention from a letter of objection. A policy objection asks the Commission to weigh interests. A methodology challenge asks it to require a stronger record before making a finding. If the FCC agrees that the coexistence evidence needs strengthening, the likely consequence is more study rather than a decision either way. For a buyer that translates straight into timeline: a methodology challenge points toward a longer proceeding.

Where the proceeding stands and what happens next

In March 2026 the FCC forwarded a draft Notice of Proposed Rulemaking on PNT to the White House Office of Management and Budget for interagency review, including the Office of Information and Regulatory Affairs. That is a real step in a long sequence. It marks progress through the process, and it is worth reading as exactly that much.

A draft NPRM in interagency review means a proposal exists in a form the Commission is prepared to circulate. An NPRM can propose alternatives, ask questions, or seek comment on a narrower change than the one requested. As of the most recent public record cited here — NextNav’s own second-quarter earnings call on 11 August 2026 — the draft remained in interagency review and the formal rulemaking was still to be published.

The remaining shape of the process, if it proceeds, is well established:

NPRM adopted and released, setting out what the Commission actually proposes and the questions it wants answered.
Comment cycle, then reply comments. On the original petition, comments closed 5 September 2024 with replies due 20 September 2024; a full NPRM cycle in a docket this contested would attract a far larger record.
Report and Order, if the Commission decides to act at all.
Effective dates and transition rules — which is where equipment questions finally live.

That last stage is where anything could be asked of installed hardware, and even there the questions that set the cost are transition questions rather than allocation questions. Would existing certified equipment be grandfathered, and for how long? Would grandfathering attach to the unit, the certification, or the deployment? Would new equipment authorisations be constrained before deployed equipment is? Would there be a coordination or protection mechanism rather than a hard cutover? All of that remains open, and years of process sit between it and a purchase order signed in 2026.

The sober read for a buyer: a reader bought this year is very likely to still be doing its job next year. The probability that the band’s channel plan looks different in five to seven years is real. Those two statements point at the same procurement response — buy hardware whose channel plan is a software object, and buy from a supplier who can push a new one to a deployed fleet.

The US band today, and the band the petition proposes

Everything above reduces to one comparison, and it is the article’s actual subject: the band a US reader is built and certified against today, and the shape the petition would leave behind.

Element	The band today, under 47 CFR 15.247	The band as the petition proposes it
Spectrum available to Part 15 RFID	26 MHz, 902–928 MHz, contiguous	An 11 MHz contiguous middle at 907–918 MHz, with the allocation of that middle among the open questions
Hop plan headroom	At least 50 hopping frequencies where the 20 dB channel bandwidth is under 250 kHz, spread across the full 26 MHz; average occupancy no more than 0.4 s per frequency in any 20 s	The same hopping discipline over a narrower span, which compresses the separation available between neighbouring readers in a dense installation
Reader power ceiling	1 W maximum peak conducted output, reaching roughly 4 W EIRP with a directional antenna	The same Part 15 ceiling for readers
What sits at the band edges	Other Part 15 devices at comparable power	A 5G uplink at the bottom edge and a downlink at the top; filings put the proposed EIRP density at 3,280 W per MHz
Governing instrument	47 CFR 15.247, as certified against today	Whatever a Report and Order eventually adopts, if the Commission acts

Buyers running estates in more than one country are managing several of these pictures at once. For the band position market by market — Europe’s lower and upper allocations, the Gulf, Africa and South East Asia — our guide to RFID frequency bands by country sets out the current state of each rather than repeating it here.

India deserves a line of its own, because its rules are recent and widely misquoted. G.S.R. 853(E) of 10 December 2021 was made expressly in supersession of the 2005 rules, and it set India’s licence-free UHF allocation at 865–868 MHz — the same span as the European lower band — with interrogators at 2 W e.r.p. and tag replies at −20 dBm e.r.p. The channel plan is specific: four channels, centred at 865.7, 866.3, 866.9 and 867.5 MHz, each up to 200 kHz, with continuous transmission up to 4 s and at least 100 ms between transmissions on a channel, referencing EN 302 208. Equipment type-approved under the older rules remains valid for its life, which is why an installed estate and a new specification can legitimately look different from each other.

Set the US picture beside any other market and the practical conclusion is straightforward. A single global bill of materials is entirely achievable. The configuration is what varies by market: the radio can be one design, while the channel list, the dwell timing, the power table and the channel-use behaviour are a different set of numbers in each region — and, if the FCC proceeding lands anywhere, a further set. Specify for that at purchase and it stays a software problem for the life of the estate.

What ‘band-configurable’ means in hardware

Reader datasheets commonly state support for FCC and ETSI bands. The phrase covers two very different engineering realities, and telling them apart is a buyer’s job.

The front end is where the region lives

A reader’s transmit and receive chain is filtered, matched and tuned for a passband. Front-end filtering keeps out-of-band energy off the receiver and holds emissions inside the mask the unit was certified against. Front-end filtering is a design choice made per region, which is why regional variants exist. A proceeding that could put high-power transmissions directly against the band edges is what makes it worth asking about at purchase time rather than at commissioning.

Regional variants are standard practice, and we build both. We configure FCC-band and ETSI-band variants per order so each unit’s front end is matched to the band it will work in.

Region profiles are the software half

A region profile is a versioned object: the channel centre list, channel spacing and occupied bandwidth, hop sequence and dwell time, the minimum gap between transmissions on a channel, listen-before-talk thresholds where the region requires them, the channel-use principle, and a calibrated conducted-power table per channel. Two readers with the same model number and different profile versions are, in regulatory terms, different radios.

Here is where each parameter lives.

Parameter	Set by firmware profile	Set by hardware design and authorisation
Channel centre list and spacing (within the front-end passband)	Yes	—
Hop sequence, dwell time, inter-transmission gap	Yes	—
Listen-before-talk threshold and channel-use policy	Yes	—
Conducted power per channel, within the calibrated range	Yes	—
Front-end filter passband and selectivity	—	Yes
Power amplifier matching and harmonic filtering	—	Yes
Antenna port tuning across the band	—	Yes
The tested configuration an authorisation covers	—	Yes — extending it is a re-testing exercise

A genuinely configurable reader carries the region profile as a first-class, versioned, signed artefact with a documented update path. That is the one that absorbs a rule change as a firmware push.

Six lines to add to a 2026 US reader purchase specification

These are worth writing into the order document, not the email thread. Each one is answerable by any competent manufacturer, and each one is expensive to retrofit.

Region profile stated per unit, by identifier. Ask for the profile name and version shipped on each serial number, rather than the country. A configuration is named like this: “Profile US-FCC-15.247, v3.1, 50-channel hop”. Require it on the packing documentation so an audit two years out can be answered from the paperwork.
Firmware-updatable channel plan, with a stated distribution mechanism. State that the channel plan must be updatable by firmware without hardware return, and require the supplier to describe how an update reaches 400 readers behind a customer firewall — staged rollout, rollback, and confirmation that a unit took the update. A fleet you can address is a fleet you update in place.
Documented transmit-power range per band and per power source. Ask for the calibrated conducted-power range in each region profile, and for the achievable range on each supported power input — a reader budgeted differently on PoE than on external DC will behave differently on a dock door than it did in the demo.
A written channel-plan change procedure, and who may run it. Regulatory configuration should be a privileged operation. Every change of channel plan should carry a named operator, an authentication step and an audit entry.
An evidence pack per destination market, listed as documents. Name the actual files you expect for each country you ship to — test reports, authorisation certificates, declarations, label artwork. Listing them as deliverables at PO stage is how they arrive before the shipment does. Our manufacturing and export documentation is assembled this way per destination.
A support window that outlasts the proceeding. Ask for a stated firmware support period and a commitment that a regulatory-driven profile update is issued within that window. If a rulemaking cycle plausibly runs into 2028 or beyond, a two-year support horizon on a seven-year asset is worth extending at purchase time, when it is free to ask for.

Questions to ask a reader manufacturer while the proceeding is open

Four questions, and the answers are diagnostic.

How are region profiles versioned and signed?

You want a version string you can read back from a deployed unit, and cryptographic signing so only an authorised image can replace a profile. A supplier who can read back the profile version on the reader in front of you can do the same across a fleet of 2,000.

What is the update path for readers already on a customer network?

Ask specifically: does an update arrive by physical access, a local utility, a vendor cloud connection, or a management server the customer controls? Many industrial estates require updates to arrive through a management server the customer controls, and the answer determines whether a future profile change is an afternoon or a project. Fleet-wide firmware and configuration management is what ReaderSense Edge MDM exists to do — and whatever platform you use, this capability is what turns a regulatory change from a capital event into a maintenance window.

What happens to an in-flight order if the allocation changes mid-production?

In our own production, lead times on configured hardware run weeks to months. Agree in advance whether units already built get updated before shipment, at what cost, and who carries it. This is a one-paragraph clause that is trivial to add now and contentious to negotiate later.

How does the supplier track and communicate regulatory change?

Ask who follows the dockets, what triggers a customer notification, and through what channel. A manufacturer that writes its own firmware can answer this concretely, because the same team that reads the rule writes the profile. We build both the hardware and the software in-house, which is the reason a change in a channel plan is a development task here rather than a supplier enquiry.

What buyers outside the US should take from this

If you are deploying in Europe, the Gulf, Africa, South East Asia or India, the FCC docket sits outside your licence — the lesson still lands.

The durable point is that channel plans have become moving objects. Europe has been adding an upper band at 915–921 MHz alongside the long-established 865–868 MHz allocation, with national adoption arriving at different times in different member states. India moved its own allocation and channel plan with G.S.R. 853(E) in December 2021, and equipment certified under the previous rules remains valid for its life — which means a mature Indian estate can legitimately hold two generations of configuration at once. Every one of those transitions asks the same operational question: can you change the radio configuration of a deployed fleet without touching it?

That capability is worth specifying once. It is free to specify at purchase, and at every future rule change it is the difference between a scheduled firmware push and a truck roll to every site. In a warehouse deployment with 60 dock-door portals, or a multi-store retail estate with ceiling readers across dozens of locations, the arithmetic settles it.

So: watch the docket, because it is genuinely unresolved and it concerns the band a large part of the world’s RFID runs on. Act ahead of it. Write the six specification lines into the next purchase order, ask the four questions of whoever is quoting, and the outcome of the proceeding becomes a maintenance item rather than a capital one — which is the position worth holding while a rulemaking is open.

Frequently asked questions

Is the FCC changing the 902-928 MHz RFID band?

As of August 2026 the band stands unchanged. A petition to restructure it has been before the FCC since April 2024 (RM-11989 / WT 24-240), and in March 2026 a draft Notice of Proposed Rulemaking on PNT went to the Office of Management and Budget for interagency review. NextNav’s own second-quarter earnings call on 11 August 2026 confirmed the draft was still in that review. A draft NPRM in review is a procedural step; a full cycle of NPRM, comments, reply comments and a Report and Order would still have to run before anything could be required of equipment.

What is NextNav's petition and how does it affect RAIN RFID?

NextNav asked the FCC to reconfigure the lower 900 MHz band to host a terrestrial 5G-based positioning, navigation and timing service as a complement and backup to GPS. The proposed geometry is a 5 MHz uplink at 902–907 MHz and a 10 MHz downlink at 918–928 MHz — 15 of the band’s 26 MHz, taken from both edges. RAIN RFID readers in the US hop across the full 26 MHz under 47 CFR 15.247, so both the usable span and the separation available to dense reader installations are what industry filings are contesting. The RAIN Alliance, GS1 US and AIM Global filed a joint response on 15 May 2026 arguing the coexistence evidence presented was insufficient.

Will existing US UHF RFID readers stop working if the band is reconfigured?

Installed equipment continues to operate under the rules it was certified against. If an order were eventually adopted, the questions that would decide the impact are transition questions: whether existing certified equipment is grandfathered, for how long, and whether grandfathering attaches to the unit, the certification or the deployment. All of that remains open. The practical response is to keep buying, and to buy readers whose channel plan can be updated by firmware, confirming that the supplier can distribute an update to a fleet already on a customer network.

Can a UHF RFID reader be reconfigured for a different frequency band?

Partly, and the split matters. Firmware can set the channel centre list and spacing, hop sequence and dwell time, minimum gap between transmissions, listen-before-talk thresholds, channel-use policy and the conducted power per channel — all within the front end’s passband and the unit’s calibrated range. Hardware design and the equipment authorisation set the front-end filter passband and selectivity, power amplifier matching, harmonic filtering, antenna port tuning and the tested configuration the authorisation covers. So a reader moves within and sometimes between regions by software, and the front end and certification scope set how far.

What frequency does the United States use for UHF RFID?

902–928 MHz, under 47 CFR 15.247. The ceiling is 1 W maximum peak conducted output, reaching roughly 4 W EIRP with a directional antenna. Frequency-hopping systems whose 20 dB channel bandwidth is under 250 kHz use at least 50 hopping frequencies, with average occupancy of any one frequency no greater than 0.4 seconds in a 20-second period.

How do FCC-band and ETSI-band reader variants differ?

The FCC variant is built and tuned for 902–928 MHz with a 50-channel hop plan; the ETSI variant for 865–868 MHz at 2 W e.r.p. under EN 302 208, with that standard’s channel plan and channel-use conditions, and a separate upper-band plan at 915–921 MHz where adopted. The differences are both physical — front-end filtering, PA matching, antenna tuning — and configurational, in the region profile that carries the channel list, dwell timing, gap discipline and power table. Identium configures FCC-band and ETSI-band variants per order.

What should I put in a US reader specification while the proceeding is open?

Six lines. State the region profile per unit by identifier and version rather than by country name. Require a firmware-updatable channel plan and a described mechanism for reaching deployed readers behind a customer firewall. Ask for documented transmit-power range per band and per power source. Require a written channel-plan change procedure with a named privileged role and an audit trail. List the evidence pack per destination market as named documents. And specify a firmware support window that outlasts the likely length of the proceeding, with a commitment that regulatory-driven profile updates are issued inside it.