Why has T-Mobile withdrawn (official) support for neutral-host deployments in CBRS spectrum? Has the CBRS brigade failed, and is its move justified, or is it just a classic carrier move by the self-proclaimed “un-carrier” on the US telco scene? And what does it mean for in-building coverage in the US if a neutral host only supports one carrier?
In sum – what to know:
Technical issues – the withdrawal of T-Mobile’s neutral-host support in CBRS spectrum is down to better performance in licensed bands, it says; others say the issues are with its own outdoor 5G SA upgrades.
Hatchet jobs – there is an admission, perhaps, that some integrators have blotted the CBRS copybook with cut-and-shut deployments; but the severity of T-Mobile’s action also suggests another agenda…
Carrier agenda – public 5G SA infrastructure is complex to map around indoor (4G) CBRS ‘islands’; cheaper multi-band 5G solutions in 2026 offer another way to bring in-building coverage into licensed bands.
Suddenly, out of nowhere, the whole CBRS experiment looks vulnerable. T-Mobile’s decision, some weeks ago, to withdraw official / unofficial support for neutral-host 4G and 5G systems in the so-called ‘innovation band’ follows on the heels of the US president’s second-term tax-and-spending (‘OBBB’) bill, which puts CBRS in the crosshairs of a new $85 billion target to find, clear, and auction up to 800 megahertz of prime spectrum. We dealt with the existential threat from the Trump bombshell before; but T-Mobile’s antics need some consideration.
Official or unofficial? The public line from T-Mobile is that it has directed its neutral-host vendors to “prioritize” its licensed mid-band spectrum instead – mostly at 2.5 GHz, plus some at 1.9 GHz – for new shared infrastructure projects. Otherwise, it will continue to offer selective support for CBRS deployments, it seems. It reiterated the standard message about the superiority of carrier-licensed spectrum, versus shared frequency bands like CBRS, on the grounds of better network performance. CBRS should be considered as a “supplemental tool”, it said.
There is irony here, of course; the firm that portrays itself as the anti-establishment ‘un-carrier’, bent on disrupting the stuffy aristocrats of US telecoms, has just adopted the most familiar establishment position imaginable, reinforcing the old orthodoxy that only licensed bands will do, and that the traditional mobile-market lock-in model delivers competition – which begets performance and innovation, they say. The irony goes further, too: T-Mobile defined the whole venue-pays / bring-your-own-coverage model, which has underpinned the neutral-host CBRS movement.
AT&T has run with it since, and, following T-Mobile’s late CBRS reversal, stands as the only US operator to offer full support for third parties looking to fund neutral-host systems in enterprise venues – and, thereby, to get a free pass to extend patchy indoor coverage. Whereas T-Mobile, plucky agitator brand, has just bunny-hopped Verizon as the most anti-CBRS carrier in the market – reckon some commentators. And triple-irony: T-Mobile could do with all the help it can get in the enterprise space, which it had been getting until now from the busy CBRS brigade.
A source says: “It is a consumer brand; it has a challenge to disrupt AT&T and Verizon in the enterprise market. They have the mindshare, historically. T-Mobile doesn’t have the profile or experience. But it was winning SIM business because building owners were offering indoor coverage on T-Mobile. It was heading in that direction.” Plans change; T-Mobile has ostensibly withdrawn from the low-cost (mostly-4G) DAS replacement market, and requested its CBRS partners to put indoor venues onto 5G small cells in licensed bands instead, under the same venue-pays terms.
Which means expensive multi-band radios from premium vendor brands, like Ericsson’s indoor Radio Dot System. Airspan’s developing AirVelocity range is a contender, as well; its purchase of Corning Wireless is interesting, in the context, and might be discussed in a separate article. But official or unofficial? “T-Mobile is not daft,” says a source. “Same as Verizon; they won’t turn business away.” The sense is that T-Mobile will not quit its existing CBRS setups – provided they are right-sized and rightly-managed. The flagship deployment with Meta is the prime example.
Money talks
The Meta deployment, covering 20 million square feet (1.8 million square metres) across 40 offices, effectively birthed the whole market. Joel Lindholm, co-founder and chief executive at CBRS neutral-host operator InfiniG, was in charge of the rollout at Meta at the time. “MOCN was the tool, eNodeBs were the radios, and MNOs were the connectivity. It was a fairly simple system, and we went through the testing with all three MNOs; all the regulatory stuff with all of them,” he says. The point is that Verizon, the original hold-out, participated at Meta.
Verizon is lined up for new CBRS deployments, too – allegedly. Whatever T-Mobile says about “supplemental support”, it will come to the table if the business is right. Lindholm says: “In a way, nothing has changed. If buildings could have been covered with licensed spectrum, they would have been. But they weren’t, and they are still starved of coverage. That doesn’t change; enterprises still need all three MNOs. But they have a choice. If the best approach is a licensed network, then okay. But we believe there are clear benefits going with the CBRS approach.”
James Estes, chief executive at neutral-host operator IONX (formerly Dense Air) says the same, more or less. “It makes it more difficult because we are going to have to work on a case-by-case basis. But in practice, it is not that different from what went before. Because everything has to get approved anyway. It’s a very tight and defined process. But T-Mobile is not actually saying you can’t do CBRS anymore; it is just saying that this is the direction it wants to go. It wants to understand our 5G evolution path, and see that it matches its own.”
In other words, there is total business pragmatism from T-Mobile, like from Verizon; but both, like AT&T at some point, want the neutral-host supply market to utilise their expensive 5G assets – and specifically their new-fangled standalone 5G (5G SA) infrastructure. Which is why, also, T-Mobile has just launched a new 5G slicing service for enterprises, hinged on its new macro SA rollout, and headlined at launch by deals with Delta Air Lines and Axis Energy Services. But slicing is a separate enterprise product for separate enterprise applications.
Indoor neutral-host coverage, or just indoor coverage, is a localised mobile broadband patchwork, which has to mesh with public infrastructure outside the building. Which is the killer for T-Mobile, perhaps; its press statement about the matter says: “Licensed spectrum delivers the certainty, performance, and scale required for 5G, while CBRS – because of its unlicensed nature, uncertain availability, and limited support in the 5G roadmap.” Rising traffic on 5G networks makes such distinctions sharper, it argues.
Like it or not, the dice is loaded – argues another source, also selling CBRS systems. More fool you, he argues, if you didn’t read the rules of the game. “You see the strategy, right? If you’ve invested billions, and have the big vendors all tied up, then why would you support a secondary market? The operators want to monetize their assets because they are in this for billions. And they swing the biggest sticks. So you’ve got to play nice, in the end – even if they are hard to work with. At a certain point, this was always going to get shut off. Because it is counter to what they want.”
He adds: “The gamble for lots of these companies was that they would build so much momentum that the carriers couldn’t turn it off – that they had two, and that they would get three. Well, guess what? Now they’ve only got one.” It’s a harsh take for a community that has, arguably, led the whole world on smart innovation in shared spectrum, providing a model for neutral-host and also for private networks (another separate discussion). He rattles off a list of blue-chip hazards in this new wild-west, in the spectral prairies between 3.55 GHz and 3.7 GHz.
Rap list
He presents them as “issues” for neutral-host operators in the CBRS band, cascading upwards into issues for the carriers on their systems. “One: the moment you offer a public network service, you trigger certain regulatory items – and, basically, you’re in the life-safety game. Two: the moment you turn it on, you lose control of your network – because everyone on AT&T and T-Mobile is going to pile onto your infrastructure and eat up your capacity. You can’t whitelist users, because then it is not public. It’s not a controlled public network; it’s a public network.
“Three: you have to maintain carrier KPIs – as a non-telco, not in the business of carrier KPIs, not making any money off the network. Four: the company providing the gateway is probably five-guys-and-a-dog. If they run out of money, then how do you explain that to the carriers? You’re effectively buying a life/safety support application that is going to spiral in complexity – which is back-stopped by a startup. It is the full lifecycle management; you have to test, certify, deliver, and then manage. Which is where the carriers have gotten twitchy.”
He adds: “In the end, the juice isn’t worth the squeeze. There is a sense their big assets are being devalued. That is where this has come home to roost.” Again, it is tough stuff – a grim and rather depressing reality check, where the establishment wins, again, and the innovators and agitators go unrewarded. But is it true? Here is a (rather measured) response from a prime CBRS operative, who prefers to remain anonymous (such is the sensitivity). “We hope this leads to a deeper analysis of the concerns raised about the quality of the customer experience delivered.”
He explains: “Is the issue really CBRS spectrum, or the maturity of solutions? If the customer experience suffers because of a fragmented architecture stitched together by an SI with RAN and MOCN from different vendors, for example, then that is a fair concern for operators. [But] it is important that this nuance is not lost, and that we don’t give up on the potential of CBRS-based neutral-host [systems] by misreading it.” It is measured, but it is also telling: first, that the charges about network performance might stand; and second, that they are probably misdirected.
In other words: CBRS spectrum is a powerful resource for enterprises and operators, and there are old pros selling carrier-grade coverage, and stick-up merchants doing hatchet jobs. Carriers might do better to weed their channels.
Others are less interested in towing a line. Lindholm – who designed the original Meta project, and founded InfiniG to replicate its supply into other venues – says, flatly, that the CBRS rap sheet is wrong. The charges, as they are presented / interpreted, are made-up, he suggests. “It is an argument that is 100 percent inaccurate,” he says about the idea that CBRS-based neutral-host systems are not geared for carrier-grade regulatory compliance, specifically about the FCC obligation for public network operators to complete 911 calls and other lawful requirements.
“Operators are expected to complete 911 calls regardless of whether their network is there or not. The FCC says you shall deliver 911 services. When we put a network in, we establish a better path to do that. We test with all the MNOs to make sure that call goes to the proper public service answering point (PSAP) – to ensure the address is correct, and the service can be delivered. We actually help them out. It is a fundamental piece. But better than that, a neutral host system provides the X, Y, and Z access location for those 911 calls.”
Neutral-hosts systems, strung-up on small cells in multi-storey buildings, provide a vertical axis read-out as well – whereas traditional carrier-sponsored DAS setups in licensed bands only provide a dispatchable location to the head-end radio in two dimensions (“to the front door of a 50-storey building”). Lindhom says: “The dispatcher has no idea which floor the call is from. But we have radios on all of them, so they know it’s the northeast corner of the 32nd floor. We solve the indoor 911 problem, and also give a more accurate location. So we help the first responders and the mobile operators.”
Outdoor pursuits
He’s not having any of the guff about brand damage, either; and his argument is very clear. “If AT&T has coverage inside a building, and T-Mobile and Verizon don’t, then whose brand is better? Very large-scale and very high-maintenance customers have a very good experience with indoor CBRS. Properly done, it is actually a way better network,” he says. “All of our time goes on that indoor piece,” he adds, as if to make the point that big mobile operators in charge of big outdoor networks are not so focused on in-building coverage.
There are five million buildings “out there”, says Lindholm, and their best chance of indoor coverage is with affordable – and totally reliable, he insists – 4G or 5G based networks in CBRS spectrum. “We are a boon to them – versus making their subscribers use Wi-Fi.” He also disputes – categorically – that (indoor) performance is an issue with CBRS. “When we put a network inside the building, we are using a different frequency to all the operators outside. So there is no quote-unquote interference. It is a lot cleaner. We are not juggling 10 bands, like outside the building.”
In fact, big customers with tough requirements have 10-times more indoor capacity with CBRS than with rival DAS systems in downtown high rises, he argues – plus round-the-clock support, and total focus. Lindholm adds: “Where some perceive there to be interference with CBRS is outside the building – where neighboring sites interfere with each other, and where there is maybe a point. Maybe. But that is outside; inside, it is excellent.” But maybe that’s the kicker for the carriers, prepared to hobble part of the CBRS market to obsess over their 5G SA networks outdoors.
Or maybe this idea of a wild west holds some water; that not all CBRS suppliers are as good as InfiniG or IONX, or whoever, and that the carriers are calling time on this brave CBRS experiment as a result. Or maybe the real story is about the gradual dismantling of a clever startup ecosystem in the so-called “innovation band,” triggered not by technical failure but by outside engineers tweaking radio parameters in 5G SA macro infrastructure, with minimal collaboration, frustrated enough to seek tighter control of the third-party systems they once greenlit.
Here is another well-placed source: “The challenge, especially as they roll out 5G SA in the public network, is ensuring smooth handoffs into and out of the CBRS network. With all these upgrades going on right now, making the in-building system feel like an integrated part of the network is not trivial. They don’t want it to act as an island – they need to be able to manage the network around it. So if changes are made on the macro side, they have to ensure it still hands off cleanly, without interference. That’s the key reason they’re moving toward licensed spectrum: control.”
The thing, as well, is that T-Mobile et al will be familiar with the 5G roadmaps from the network vendors, and know there are cheaper – comparably-priced, reckon some sources – neutral-host solutions coming in 2026 with multi-radio support for all of their licensed spectrum holdings in the US. The CBRS crowd knows it too; the slightly jarring aspect, perhaps, is how quickly some of them have switched horses – to let their old CBRS nags whinny and neigh, as they set a new course on licensed steed, in red, blue, and magenta. But maybe that’s just good business.
