CTIA and some of its mobile-industry members continue to push policy messaging and commissioned studies claiming that the US has a significant “gap” in mid-band commercial spectrum compared to other countries, especially in China and Europe, which undermines US 5G and future 6G networks.
There are two problems, however, with these doom-mongering claims of a mid-band deficit:
- CTIA misrepresents “mid-band” as frequencies from 3GHz upwards. The rest of the mobile industry refers to mid-band as starting at 1GHz.
- CTIA’s international comparisons mischaracterize some other countries’ low-power, shared mid-band spectrum allocations as high-power exclusive bands.
Together, these numerical mis-steps mean that CTIA’s advocacy systematically undercounts the US availability of mid-band frequencies, and overcounts the rest of the world. When comparing apples to apples, the claimed mid-band spectrum deficit in the US shrinks hugely or disappears.
Mid-band starts at 1GHz, not 3GHz
CTIA uses an unusual, arbitrary and self-serving definition of “mid-band” to make it appear that US cellular operators have less spectrum access. Together with its consultants such as Accenture, it has unilaterally defined spectrum ranges as:
- Low-band (0.3 – 3GHz)
- Lower mid-band (3 – 8.4GHz)
- Upper mid-band (8.4 – 24GHz)
- High-Band (>24GHz)
Creating and changing its own convenient definitions is not new. In 2016, CTIA called anything above 6GHz “high band”. In 2018, reports CTIA commissioned from external consultants started referring to mid-band between 3-24GHz. But in 2020 it described mid-band as stretching from 3-7GHz, with low-band below 3GHz. But by 2022 it had adjusted this definition yet again to include frequencies up to 8.4GHz, as it became clear that might conceivably be used for cellular at some point.
Meanwhile, pretty much everyone else in the mobile industry defines mid-band as 1-6GHz or 1-7GHz, including GSMA, 5G Americas, Nokia, Ericsson and many more.
- GSMA, 2022: “Mid-bands are found at 1-7 GHz”
- CTIA member T-Mobile: Refers to mid-band as 1-6GHz in a recent blog, and specifically identifies its 2.5GHz range as mid-band
- 5G Americas, 2023: “mid-band, which ranges between 1 GHz and 6 GHz”
- Nokia: “Mid-band spectrum (1 GHz – 6 GHz) is considered perfect for 5G”
- Ericsson, 2022: “Network professionals talk about this area in terms of sub-6 GHz… The spectrum used here is in the 1 GHz to 6 GHz range”. On an industry analyst call in December 2024, the company confirmed that it views 1-7GHz as midband, with 5G deployments mostly using 2.3GHz upwards.
- GSA (the Global Mobile Suppliers’ industry) released a report in July 2024 which discussed actual deployed mid-band networks and frequency ranges, between 1.71GHz and 5GHz
- Even CTIA consultants AnalysysMason, in a 2022 report commissioned by CTIA, noted (perhaps with some embarrassment) that “The 2.5GHz band … shares many of the same characteristics of the identified mid-band spectrum … and could therefore properly be grouped with mid-band spectrum as well.”
In other words, the industry consensus agrees that 1-3GHz is part of the mid-band. In early discussions about 6G spectrum, some have started to refer to “upper mid-band” between 7-24GHz, but they still refer to lower mid-band from 1GHz and up.
The US telecom industry’s own standards body ATIS (Alliance for Telecommunications Industry Solutions) holds the sensible view that low / mid / high band terminology is too vague, saying that “in general, it is recommended that the above terms are acceptable only if ambiguity can be avoided”.
Why does this matter? It means that CTIA can artificially exclude its members significant AWS licenses in the 1-3GHz band and T-Mobile’s huge 2.5GHz holdings from its arbitrary count of “mid-band” availability, even as T-Mobile itself regularly refers to those licenses as “mid-band”.
CTIA also conveniently omits any references to 4G / 5G cellular being able to use unlicensed spectrum bands such as 5GHz for extra localized capacity, or indoor coverage. Technologies such as 4G LTE-U (Unlicensed), MuLTEfire, 5G NR-U and others are well-understood and have been used by some US operators in the past. If they really needed extra spectrum, they’d make use of this optional extra resource, at least opportunistically.
If CTIA instead used the same definition of mid-band used by everyone else, their claims look very different. Taken together, AT&T, Verizon, T-Mobile and Dish/EchoStar have around 870 megahertz of true midband spectrum. This is much closer to other global peers, including China and various European countries.
International comparisons
Similar inaccuracies or questionable assertions occur in CTIA’s claims and commissioned research centering on comparisons with other markets, especially China, but also Europe.
For instance, very few countries have allocated the 3.3-3.4GHz range to MNOs, with even fewer actively using it for full-power exclusive use. In China, this band is specifically allocated for low-power indoor use, via network-sharing arrangements between China Unicom, China Telecom & China Broadcasting Network. This is in stark contrast to how CTIA characterizes its use.
Indeed, when CTIA commissioned AnalysysMason to look at 3-7GHz mid-band allocations for mobile in 2022, it did not distinguish between high-power nationwide licenses and low/mid-power licenses for shared local or private network use. AnalysysMason noted “totals include spectrum that is available in several countries for private 5G use. In particular, several regulators in Europe have set aside spectrum specifically for private 5G networks”. Yet subsequent comments by CTIA have overlooked that caveat, and just assumed it was all equivalent to auctioned, exclusive use bands, which is most definitely not the case.
CTIA member AT&T has also separately claimed in its recent CBRS NPRM comments that “many countries have assigned or planned to assign a large, contiguous swath of 3 GHz spectrum for mobile broadband, including Japan (3.4-4.1GHz), the United Kingdom (3.4-4.2GHz), France (3.4-4.0GHz), South Korea (3.4-4.0GHz), and Australia (3.4-4.0GHz)”.
This is wrong for several countries – again, it is combining normal exclusive, national high-power assignments with local shared-spectrum allocations for private networks and community FWA networks. For instance in the UK, 3.4-3.8GHz is allocated for national MNO use, but 3.8-4.2GHz falls under the “Shared Access Licence” framework for low/medium power deployments on a location-by-location basis. In France, the 3.8-4.0GHz range is currently used by about 90 private networks on a trial basis. Australia has local “Area Wide Licenses” in the 3.75-3.95GHz range, for enterprises such as mining and remote-area community FWA broadband.
In other words, CTIA and AT&T are ironically and unwittingly advocating for some good policy positions here; a focus by governments on improving indoor coverage via network-sharing and neutral host models is very wise. And the availability of up to 400 megahertz of local, shared spectrum for private 5G, specialist providers, or community FWA in some markets is significantly more bandwidth than that available from CBRS in the US.
Conclusion
CTIA’s argument that the US lacks sufficient mid-band spectrum for mobile operators’ 5G networks is deeply flawed. It undercounts existing US allocations, especially in the 2.5GHz range that its own member T-Mobile often identifies as mid-band. And its international comparisons incorrectly include a variety of low-power and shared spectrum when adding up other countries’ allocations to exclusive high-power licenses.
In summary, it seems that CTIA needs to revisit its definitions, assumptions and comparisons. It will find that the US has plenty of mid-band spectrum, even if it is currently distributed unevenly between its members.
