Will Phone Carriers Disappear Because of Starlink Mobile?

Two joint ventures in one year. On October 1, 2026, AT&T, T-Mobile and Verizon signed the definitive agreement for a satellite venture that pools their spectrum, after each CEO had refused, on the record, to host SpaceX as a virtual operator. On October 19, Deutsche Telekom, Orange, Telefónica and Vodafone signed a letter of intent to bid together for Europe’s satellite frequencies. Carriers that spend their lives fighting each other do not join hands over a gadget. They do it over a threat.

So the clickbait question deserves a serious answer. The honest one is neither “yes” nor “no”: it is “which part of a carrier, and when”. This article takes a carrier apart into its four layers, checks which layer Starlink Mobile actually reaches at each horizon, tests the comparison with earlier disruptions, and ends with what none of it changes for a company choosing a phone system today. The deal mechanics, quotes and FCC conditions are covered in detail in Starlink Mobile vs AT&T, Verizon and T-Mobile; this piece does not repeat them.

What a carrier is actually made of

“Carrier” sounds like one thing. It is four businesses stacked on top of each other, and they do not share the same vulnerability.

Layer one: spectrum. Licensed frequencies are the scarce input, granted by a regulator, country by country. Without them you can launch satellites but not sell service. SpaceX now holds about 65 MHz of exclusive nationwide US spectrum. Everywhere else it borrows from a partner carrier.

Layer two: the access network. Towers, small cells, fibre to the tower, and now satellites. This is the only layer where a satellite competes with a carrier today, and the only one where orbital physics argues it will ever compete fully: a beam that covers hundreds of square miles serves far fewer users than a grid of cells two miles apart, which is why Verizon’s CEO put the number at 10 to 40 homes per square mile versus 500 to 2,000 for fixed wireless.

Layer three: the core network and numbers. Number allocation, emergency calling, interconnection with every other network, number portability, roaming agreements, lawful intercept. Each requires a licence and an operational core in the country concerned. This is the slowest layer to conquer and the best protected by law. In today’s direct-to-cell service the traffic lands in the partner carrier’s core network: the phone numbers, emergency routing and interconnection still belong to the carrier.

Layer four: the customer relationship. The brand, the bill, the stores, the support line, the bundle with the home internet. Starlink has a brand and 12 million broadband customers. It has no mobile customers of its own, anywhere.

Keep this stack in mind. “Carriers disappear” means Starlink takes all four layers in dozens of countries with different laws. “Carriers become dumb pipes” means it takes enough of layer two that layers three and four lose their pricing power. These are very different claims, and the evidence supports only the second one, and only partially.

Short term, 2026-2027: a safety net sold by the carriers themselves

What is certain. Starlink Mobile exists as a wholesale product. SpaceX said at MWC in March 2026 that about 650 direct-to-cell satellites were active and that the service had 10 million monthly users. Carriers sell it: T-Mobile in the US since July 2025, One NZ in New Zealand, Telstra and Optus in Australia, Rogers in Canada, KDDI in Japan, Entel in Chile and Peru, Kyivstar in Ukraine since November 2025, Virgin Media O2 in the UK since February 2026, with Airtel Africa signed in December 2025 for 14 markets. In every case the carrier keeps the customer, lends its own frequencies and bills the add-on. O2 charges £3 a month and restricts it to Samsung Galaxy S25 phones. Kyivstar gives it away and reported nearly 5 million users out of 22.4 million subscribers in March 2026.

What those users get is narrow: text, location and a short list of apps, outdoors. In rural New York drive tests, 60% of text attempts succeeded with an average round trip of 1 minute 17 seconds. Ookla counted 0.46% of US Speedtest users with a direct-to-device connection in March 2026, and the number fell when T-Mobile started charging $10 a month at the end of free trials. Why a normal phone cannot yet make a regular call through a satellite is explained in how direct-to-cell works.

What is announced. The V2 constellation: about 1,200 next-generation satellites scheduled to launch in 2027, with upgraded service targeted for the end of that year. They will transmit on SpaceX’s own satellite bands rather than only on borrowed carrier spectrum.

What is a bet. A retail Starlink phone plan anywhere before 2028. SpaceX has stated the intent for the US and nothing else: no price, no plan structure, no handset list.

Medium term, 2027-2029: the three gates between supplier and competitor

This is where the question is decided, and it is decided by three gates that have nothing to do with rockets.

Gate one: spectrum, and Europe’s reallocation

In the US, SpaceX passed the gate when the FCC approved the EchoStar transfer on May 12, 2026, with permission to use the band from space, on the ground or both. In Europe, it has not. The 2 GHz band, harmonised across the EU since 2008, is held by EchoStar and Viasat in two 2x15 MHz blocks that expire in May 2027, and EchoStar’s rights are due to pass to SpaceX under the US deal.

On May 27, 2026 the European Commission proposed to reallocate the band through an EU-wide comparative selection: one third for government use tied to the IRIS² constellation, and of the commercial two thirds, half reserved for EU-based new entrants and half open to EU and non-EU operators, with a cap of 2x10 MHz per operator and non-EU players limited to one third of the frequencies. Existing licences could be extended two years for current services without the right to a full new direct-to-device offer. New rights are not expected before 2029, and Parliament and Council have yet to vote.

Aetha Consulting’s Amit Nagpal judged that “Starlink doesn’t come out of this too badly”, since a beauty contest favours a deep-pocketed operator with a track record. The politics are rougher: FCC Chairman Brendan Carr said “reciprocity will continue to be the touchstone”, and the US ambassador to the EU warned, according to the Financial Times, of “devastating” retaliation if Brussels restricted US companies. The four-carrier letter of intent was signed the next day, framed around the Commission’s sovereignty goal. Telefónica’s Jesus Millon Millan had summed up the sector’s mood in Paris on September 16: “It’s complicated.”

Gate two: handsets

A phone modem only transmits on the bands its radio front end was built for. Today’s services work on most recent phones precisely because they use carrier frequencies those phones already support. V2 satellites will use SpaceX’s satellite bands, which almost no handset supports. SpaceX is developing a modem for chipmakers, showed a MediaTek test device at MWC 2026, and Deutsche Telekom, the only major European carrier with a dated Starlink plan, says the number of compatible devices in Europe must grow before its 2028 launch. Elon Musk himself has put chip adaptation at around two years. A phone plan nobody’s phone can use is not a competitor to anyone.

Gate three: the ground

The physics argument the carriers make is correct for cities and wrong for the countryside. SpaceX’s answer, stated on the August 4, 2026 call, is “a large number of small stations”: femtocells housed with customers’ Starlink dishes, deployed incrementally, using the EchoStar bands on the ground. That is a conventional cell site at 2 GHz, able to reach indoors across a neighbourhood. It is also a tower business, with rooftops, power, permits and backhaul, and it exists today only as a stated intent with no hardware details. Where Starlink dishes are dense, which skews rural, it could work. In a dense suburb it depends on a density figure SpaceX has not published.

Long term, 2030 and after: disappear, or become the pipe?

Beyond 2030 there are no facts, only structures and precedents. Three precedents are usually cited, and each teaches something different.

Cable versus streaming. Streaming did not build a network; it rode over the cable operator’s pipe and took the content relationship. The operator survived as the pipe, with weaker pricing power. Starlink is doing the reverse: it builds the pipe. For it to take the relationship, it needs layers three and four in every country, which streaming never needed. The precedent predicts pressure on margins, not disappearance.

Skype versus the voice business. VoIP over the internet hollowed out the price of a phone call, and carriers kept the subscriber by bundling voice into data plans. The lesson is that a carrier survives the loss of a product if it keeps the number and the bill. That is why layer three matters more than layer two.

The MVNO. A virtual operator rents the access layer and owns the customer. It is the fastest route to an indoor, urban Starlink plan, and the one all three US carriers closed in 2026 in nearly identical words. The refusal tells you where they think the danger is: not in the sky, but in a Starlink-branded SIM in a city.

Put the three together and the realistic long-term scenario is not a carrier graveyard. It is a world where the access layer is partly interchangeable, satellites and rooftop cells cover what towers never reached, and the players that own spectrum, the phone and the app negotiate from strength. Roger Entner of Recon Analytics described the purpose of the US joint venture as making satellite providers “more interchangeable”. The European letter of intent, which explicitly rules out building satellites and promises open access to all operators, is the same move: mutualise so that no single sky supplier holds the leverage.

Europe adds a public actor to the long term. IRIS², the EU’s own multi-orbit constellation under a 12-year concession signed in December 2024 with SES, Eutelsat and Hispasat, was expanded on August 7, 2026 to 348 satellites, including 66 for defence and emergency services, with first launches in 2029. It is not a direct-to-phone rival to Starlink, but one third of the 2 GHz band is earmarked for it in the Commission’s proposal, which is why Europe will not let that band be decided by the market alone.

What none of this changes for a business phone system

A company choosing its telephony in 2026 can read the whole saga and change nothing, for a structural reason: a business phone system lives above layer two. Numbers, routing, recording, CRM synchronisation, compliance and reporting come from the VoIP provider, not from the tower or the satellite carrying the packets. That decoupling is the entire point of VoIP, and it means a Starlink satellite is, from the phone system’s point of view, just another access network.

Two practical consequences follow.

VoIP over a Starlink dish works today. Published latency of 25 to 60 milliseconds sits well inside the voice budget; the real issues are satellite handovers, a narrow upload lane and router queueing, each of which has a fix. The details and the setup are in does VoIP work over Starlink. Sales teams in sites without fibre already run their dialer over it.

Direct-to-phone satellite will be a safety net for staff, not a replacement plan. When voice and data arrive on the V2 constellation, the use case is the field rep between two valleys, the on-call engineer on a site with no coverage, the lone worker. It complements the company’s mobile fleet. It does not replace the phone system, the licences or the office connection, and whether your team can already make calls with Starlink depends on which of the two Starlink paths you mean.

The signals to watch, with dates

• 2027: launch of the V2 constellation and the first handsets supporting SpaceX’s satellite bands. If either slips, every downstream date slips.
• May 2027: expiry of the EchoStar and Viasat 2 GHz licences in Europe, and the fate of the two-year extension.
• Vote on the Commission’s 2 GHz regulation by Parliament and Council, and the final size of the non-EU cap.
• End of 2027: SpaceX’s own target for upgraded service in the US; first evidence of ground cells next to dishes.
• Early 2028: Deutsche Telekom’s launch, or delay, in ten countries.
• The two joint ventures’ satellite partners: who the US venture connects, and whether the European four move from a letter of intent to a bid.
• Any US carrier reversing on wholesale access, or any filing toward an acquisition. Nothing has been filed.

Sources

• AT&T, T-Mobile and Verizon Launch Joint Venture, Name Paul Roth Interim CEO (AT&T, October 1, 2026)
• AT&T, T-Mobile and Verizon form D2D JV for rural coverage (Light Reading, May 14, 2026)
• SpaceX wants to be a fourth major network, but it’s building around towers, not on them (Wireless Estimator, August 6, 2026)
• Verizon CEO plies physics to baffle satellite broadband, declares no MVNO backdoor (SDxCentral, July 30, 2026)
• FCC approves $40B EchoStar spectrum sales to SpaceX and AT&T (SDxCentral, May 13, 2026)
• At MWC, SpaceX execs tout Starlink V2, key carrier partner for it (Urgent Communications, March 8, 2026)
• SpaceX Scheduled 2027 Launch of 1,200-Satellite Mobile Constellation (Broadband Breakfast, March 3, 2026)
• Ookla finds D2D scale is elusive despite increased traction (Mobile World Live, April 22, 2026)
• EU risks US ire with limits on 2GHz satellite spectrum (Light Reading, May 28, 2026)
• EU telcos close ranks against Starlink in bid for satellite spectrum (Light Reading, September 8, 2026)
• Direct-to-cell : la guerre des fréquences bat son plein (Next.ink, October 20, 2026)
• Les opérateurs européens s’allient face à Starlink (Journal du Geek, October 21, 2026)
• Telekom closes the last white spots in Europe with Starlink satellite-to-mobile (Deutsche Telekom, March 2, 2026)
• MasOrange to Trial Starlink Direct-to-Cell in Spain (Via Satellite, February 3, 2026)
• Virgin Media O2 switches on Starlink-powered smartphone satellite service (TechRadar, February 2026)
• O2 switches on Starlink satellite service for Samsung Galaxy S25 in UK first (ThinkBroadband, February 2026)
• Ukraine first country in Europe to get Starlink satellite phone service (The Register, November 25, 2025)
• Kyivstar plans to expand Starlink Direct to Cell in 2026 (Interfax-Ukraine, March 2026)
• Airtel Africa Partners With Starlink for Direct to Cell Service (Via Satellite, December 17, 2025)
• Telstra launches Starlink satellite-to-mobile texting (ACS Information Age, 2025)
• New Zealand launches nationwide direct-to-smartphone Starlink service (SatellitePro ME)
• Starlink Direct to Cell: partner carriers (Converge Digest)
• Starlink 5G in every phone as SpaceX develops mobile satellite chip (Notebookcheck, September 17, 2025)
• ESA confirms kick-start of IRIS² with European Commission and SpaceRISE (ESA, December 2024)
• EU IRIS²: 66 defence satellites, 2029 launches (Sovereign Magazine, August 2026)
• Starlink says SpaceX targeting 2026 for launch of Starship-ready terabit satellites (The Register, July 16, 2025)

All sources accessed October 23, 2026. Primary sources for the US spectrum deals, FCC conditions and executive quotes are listed in Starlink Mobile vs AT&T, Verizon and T-Mobile.