How T-Mobile Took the 5G Lead, and What SuperMobile Actually Buys a Business

Before renewing our own carrier contract, we ran a full procurement analysis, which almost nobody does for a phone bill. Current rate cards at all three carriers, FCC auction records, spectrum filings, outage reports, and every independent network study published in 2025 and 2026. It took the better part of a week.

The conclusion surprised the part of us that grew up believing "Verizon is the network." For business 5G in 2026, T-Mobile is the strongest overall package by the independent test results, and its flagship plan, SuperMobile, includes capabilities the other two carriers do not sell at any price. The rest of this piece walks through how that happened, in megahertz and milliseconds. It also covers where T-Mobile remains the wrong choice, because otherwise this would just be an ad.

The spectrum math

Radio spectrum behaves according to one iron law: the lower the frequency, the farther a signal travels and the better it penetrates walls, but the less data it carries. Free-space path loss grows with the square of frequency. Moving from 600 MHz up to 3.7 GHz costs a signal roughly 15.8 dB at the same distance, a factor of about 38 in raw power terms.

That law splits usable spectrum into three tiers. Low band (600 to 900 MHz) covers 5 to 15 kilometers per tower and reaches basements, but its channels are thin, often 5 to 20 MHz. Millimeter wave (24 to 47 GHz) carries enormous bandwidth but dies within a few hundred meters and is blocked by a pane of glass. The prize sits in the middle: mid-band, from roughly 2.5 to 3.7 GHz, where a carrier can get 100+ MHz channels with a cell radius of 1 to 5 kilometers on the existing tower grid.

Every difference between the three networks in 2026 traces back to who holds how much mid-band, at which exact frequency, and since when. Population-weighted, the current holdings look like this. T-Mobile fields about 185 MHz at 2.5 GHz (band n41) plus roughly 40 MHz of C-band, around 380 MHz of total sub-6 GHz depth, which works out to about 2.9 MHz per million subscribers. AT&T holds roughly 80 MHz of C-band plus 74 MHz at 3.45 GHz, about 375 MHz total and around 3.1 MHz per million. Verizon runs on 161 MHz of C-band (band n77, at 3.7 GHz) and roughly 280 to 297 MHz total, about 2.0 MHz per million subscribers.

Note that last ratio. Verizon carries the largest subscriber base on the thinnest sub-6 GHz portfolio of the three. That single number explains most of what its customers feel at a congested cell.

The Sprint dowry and the $53 billion detour

Two events explain how T-Mobile ended up with the deepest mid-band position: the Sprint merger and the C-band auction.

When T-Mobile absorbed Sprint in April 2020, the transaction was reported as a subscriber play. The engineering asset was Sprint's 2.5 GHz portfolio: roughly 160 MHz in the top 100 markets, nationwide, already licensed and cleared, with radios for it in every vendor catalog. T-Mobile lit it up within weeks of closing and reached 200 million people with mid-band 5G by November 2021. Counting the entire enterprise value of Sprint against that spectrum works out to roughly $0.52 per MHz-pop, and the price included 30 million customers, a tower grid, and Sprint's PCS holdings as a bonus.

Verizon took the other path. Having skipped the 600 MHz auction in 2017, it entered the 5G era with almost no mid-band and a heavy bet on millimeter wave, which independent measurements later showed carrying under 1 percent of user connection time. That left the 2021 C-band auction as its only nationwide exit. The bidding closed at $81.2 billion across all winners, about 170 percent above pre-auction projections. Verizon's share: $45.45 billion for the licenses, plus its portion of a $9.7 billion accelerated-clearing fund paid to the satellite operators occupying the band, bringing its all-in cost to roughly $53 billion at about $1.08 per MHz-pop. Double T-Mobile's unit price.

The money then sat in escrowed licenses while T-Mobile kept building. The first 60 MHz of C-band did not go live until January 2022, weeks late, after the FAA dispute over radar altimeters forced power limits and exclusion zones around 50 airports. The full 161 MHz average only became usable in August 2023, once the satellite incumbents finished vacating. By that date T-Mobile had been running uncontested mid-band 5G for nearly three and a half years.

Physics compounded the head start. The path-loss delta between 2.5 and 3.7 GHz is about 3.4 dB in free space and grows with clutter, which translates to roughly 25 to 30 percent more cell radius at the lower frequency. Depending on whose engineering model you trust, covering the same area at 3.7 GHz takes between 1.3 and 1.8 times as many sites. T-Mobile's mid-band therefore spread across its existing grid, while its competitors' C-band required densification. The result shows up in the coverage statistics: T-Mobile's Ultra Capacity mid-band footprint passed 300 million people in October 2023 and, per T-Mobile's own coverage disclosures, spans more mid-band square miles than both rivals combined.

The architecture gap

Holding the spectrum only matters if the core network can exploit it.

T-Mobile launched the world's first nationwide 5G standalone core in August 2020. Standalone means the 5G network signals and routes on its own, with no 4G anchor underneath. AT&T did not reach nationwide standalone until October 2025, and Verizon's migration is still under way. Five years of standalone maturity is why T-Mobile shipped voice-over-5G (VoNR) first, deployed 5G-Advanced nationwide in 2024, runs four-carrier aggregation in its mid-band, and, most relevant here, sells network slicing commercially while its competitors are still piloting it.

The measurement record reflects the stack. T-Mobile took 12 of 16 awards in Opensignal's July 2026 national report, including Reliability Experience and Consistent Quality, the metric that checks whether sessions clear the latency, jitter, and packet-loss thresholds demanding applications need; in the January 2026 edition, its reliability score of 937 led Verizon's 933 and AT&T's 932. Ookla's Speedtest data for the first half of 2026 crowned it Best Mobile Network, with a median 5G download of 314 Mbps, roughly 100 Mbps ahead of either rival, a Consistency score of 89.7 percent, and the lowest measured latency. And on the stricter 5G availability definition Opensignal introduced in January 2026, T-Mobile phones had a 5G signal 91.2 percent of the time, AT&T 88.7 percent, and Verizon 59.3 percent.

What SuperMobile is, in engineering terms

In August 2025 T-Mobile replaced its business lineup with CoreMobile, ProMobile, and SuperMobile. The first two are competent tiers with 50 GB and 200 GB of priority data respectively. SuperMobile is the one built on capabilities the standalone core makes possible.

A dedicated network slice

Don't confuse this with ordinary priority data. Cellular schedulers rank traffic by QoS class (QCI); priority data means your packets sit in a better position within the same queue as everyone else, and under true congestion even a good queue position degrades. A slice is a different construct: a virtualized partition of the network with resources reserved for it end to end, the same mechanism carriers use for first-responder services. T-Mobile states SuperMobile lines receive up to twice the network resources of a typical user, with reduced latency. During a stadium event or a downtown lunch rush, your traffic keeps a reserved lane instead of a better spot in a jammed queue. No other US carrier sells a nationwide slice on a commercial smartphone plan as of this writing.

Satellite, included

T-Satellite runs on a constellation of more than 650 Starlink direct-to-cell satellites and launched commercially in July 2025. Messaging works today, app data followed in October 2025, and voice capability is on the roadmap as the second-generation satellites launch. For vehicles that cross the roughly half of US land area where no terrestrial mid-band exists, it is a real coverage floor for those routes.

Security in the base plan

Scam Shield Premium and the Threat Protect device security service are included. Our professional opinion as a security firm: neither replaces managed endpoint protection or mobile device management, and we deploy both on top. As a default baseline for every line on an account, however, it exceeds what the equivalent AT&T and Verizon tiers bundle.

The rest of the specification: unlimited premium data with no deprioritization threshold, 300 GB of full-speed hotspot per line, 5 GB of full-speed roaming in more than 215 countries, and a five-year price guarantee on the plan rate. Two caveats on that guarantee. It covers the base rate, not taxes and surcharges. And it is best read against the field: Verizon Business still advertises a three-year guarantee on its My Biz Plan, though one that excludes taxes, fees, and its Economic Adjustment Charge and cancels if you change the plan; Verizon's consumer side dropped its price lock for new customers in June 2026; and AT&T's promotional fine print states outright that rate plan prices may increase. Five years is the longest commitment any US carrier currently puts in writing.

The rate card, without the asterisks

On multi-line business accounts, SuperMobile lists at $50 per line. Bank-account AutoPay brings it to about $46, and published promotional credits can push the effective rate near $42. The closest engineering equivalents elsewhere: Verizon's My Biz Plan needs its $10 Premium Network Experience add-on to stop being deprioritizable, landing around $39 per line before a separate Economic Adjustment Charge of roughly $3 to $4 per line that sits outside the advertised price. AT&T's Premium 3.0 runs $50 per line with bank AutoPay, includes strong hotspot and international allowances, and carries no price protection.

Three pieces of fine print apply across all three carriers and account for most of the gap between the banner price and the invoice. No current business plan includes taxes and fees in the advertised rate; that era ended industry-wide. The full AutoPay discount now requires payment from a bank account at every carrier, with credit cards earning a reduced discount or none. And promotional phones are financed through 24 to 36 months of bill credits that function as a contract in everything but name; bringing your own devices preserves the ability to renegotiate annually.

Where T-Mobile is still the wrong answer

The mid-band advantage is concentrated where people actually are. On raw geography the picture inverts: Verizon's LTE footprint covers roughly 60 percent of US land area, AT&T about 57 percent, T-Mobile about 45 percent, and rural population coverage runs 93.4, 93.2, and 79.6 percent respectively. RootMetrics, whose methodology drives every state and 6,800 indoor venues rather than sampling where subscribers cluster, still scores Verizon and AT&T ahead on that style of reliability, and Verizon has now taken its award for most reliable 5G twelve consecutive test periods.

For an operation running long rural corridors, the sound engineering answer remains a Verizon or AT&T line for those routes, or a dual-eSIM configuration that pairs a primary carrier with a low-cost second network as failover. Building interiors follow the same logic: penetration rides on low band, where Verizon and AT&T hold deep 700 and 850 MHz positions, and no plan tier fixes a warehouse that needs a site survey.

One more finding from the outage record, because no coverage map shows it: the largest availability failures of recent years were operational, not radio. AT&T's February 2024 event, which blocked 92 million calls including more than 25,000 attempts to reach 911 over 12 hours, began with a misconfigured network element pushed to production overnight without peer review, per the FCC's report. Verizon logged multi-hour national-scale events in both September 2024 and August 2025. Whichever carrier you choose, the biggest single risk to your uptime is a bad change window in someone else's core network, which is why we design failover paths instead of picking a favorite carrier.

The playbook, and the offer

We applied the same analysis to our own fleet: SuperMobile where the slice, the satellite floor, and the security baseline earn their premium, lighter tiers where they would be wasted. AutoPay from a bank account from the first billing cycle. No promotional hardware, and if you carry device financing, time any migration to the end of the installment schedule, since switching mid-installment forfeits credits. Every negotiated credit goes in the written order confirmation, because verbal commitments from carrier reps are unenforceable; only paper is.

The same rules generalize to any business. Buy the priority class rather than the logo. Match each line's tier to its job, keep your numbers portable, and assume any advertised price survives 12 to 24 months at best.

If you would rather not spend a week reconstructing carrier fine print, that is work we now do for clients. We map your fleet, your coverage geography, and your security requirements against what the carriers actually sell, then hand you the order sheet. Get in touch and we will start from your current bill.

Disclosure: EFROS has no partnership, reseller, or compensation arrangement with T-Mobile or any carrier named here. We are a paying customer, and the analysis is our own.

Sources for the network claims above: Opensignal U.S. Mobile Network Experience reports (January and July 2026), Ookla Speedtest Connectivity Report (H1 2026), RootMetrics State of the Mobile Union (2H 2025 and 1H 2026), FCC Auction 107 public notices, the FCC Public Safety and Homeland Security Bureau report on the February 2024 AT&T outage, FCC Broadband Data Collection coverage data, and carrier plan documentation current as of August 2026. Pricing verified August 2026 and subject to change.