Starlink Direct to Cell: How Satellite-to-Phone Works

By Daniel Marsh, satellite communications correspondent, Starlink News — covering constellation build-out, spectrum policy and carrier rollouts. Field testing conducted by the author in northern Nevada and eastern Oregon, June to August 2026. Reviewed by the Starlink News editorial desk. Last reviewed 19 September 2026. Sourcing rules, testing method, author credentials and published corrections are set out in How we reported and verified this article.

Starlink Direct to Cell turns SpaceX satellites into cell towers in space. An ordinary 4G LTE handset — no dish, no clip-on antenna, no special app — connects to a spacecraft roughly 550 kilometres overhead using spectrum already licensed to the user’s own mobile carrier. As of September 2026 the service is commercially live in a short list of markets, it is overwhelmingly a messaging service rather than a broadband one, and the gap between what it does today and what SpaceX has filed to build is the single most misunderstood thing about it.

What follows is a reporter’s map of the technology: the mechanism, the named carriers and regulators, the real throughput, the coverage holes, where satellite-to-phone sits alongside Starlink’s dish-based satellite internet plans, and how it compares with Globalstar, Skylo, AST SpaceMobile and a Garmin in your backpack. Where a number is not published by SpaceX, a carrier or a regulator, we say so rather than inventing precision.

Starlink Direct to Cell at a Glance

What Is Starlink Direct to Cell and How Does It Work

Definition: Starlink Direct to Cell is a satellite-to-phone service in which purpose-built Starlink satellites carry an LTE (and, on later hardware, 5G) base-station payload and broadcast in a partner mobile operator’s own licensed cellular spectrum, so unmodified smartphones register to the satellite exactly as they would to a terrestrial tower.

The critical design choice is what the satellite pretends to be. A standard Starlink satellite talks to a phased-array user terminal in Ku-band, at frequencies no phone can reach. A Direct to Cell satellite instead carries an eNodeB — a cellular base station — behind a large deployable antenna, and radiates in the low-GHz bands that mobile networks already use. Your phone does not know it is talking to something in orbit. It sees a cell, camps on it, and sends a message. SpaceX describes the architecture in its own Direct to Cell service documentation, which remains the primary source for what the payload is designed to do.

The satellite is the tower, and the tower is moving

Because the payload flies at about 550 km in low Earth orbit — the same shell as the rest of the constellation, and the reason SpaceX can add Direct to Cell birds to existing launch manifests — each satellite is only over a given point on Earth for a few minutes. Continuity therefore depends on constellation density rather than on any single spacecraft. That is why Direct to Cell only became plausible after SpaceX had thousands of satellites in orbit and a launch cadence measured in dozens of spacecraft per week; the deeper mechanics of that build-out are covered in our piece on the Starlink Satellite Constellation: Orbital Mechanics, Launch Cadence, and Deorbiting, which explains how constellation scale and the 550 km LEO altitude shape everything the service can and cannot do.

Two engineering consequences follow from that altitude, and both are matters of orbital mechanics rather than marketing. Orbital velocity at 550 km is about 7.6 kilometres per second — roughly 27,000 km/h, a figure you can derive from the altitude itself rather than take on trust. At that closing speed the carrier frequency is Doppler-shifted, and the payload has to pre-compensate so a handset radio designed for a stationary tower can still lock. The second consequence is timing advance: LTE’s protocol timers were written for towers a few kilometres away, not for a base station 550 kilometres up, so those timers have to be stretched at the satellite end. Both problems are solved in the spacecraft. That is precisely why the phone needs no modification, and it is the whole reason the service can exist on handsets sold years before the first Direct to Cell satellite launched.

Spectrum, not hardware, is the scarce resource

Direct to Cell does not use satellite spectrum. It uses terrestrial mobile spectrum that a national regulator has already granted to an operator — T-Mobile’s PCS holdings in the United States being the reference case. That is the whole commercial model: SpaceX supplies orbital infrastructure, the carrier supplies spectrum rights and the customer relationship, and the regulator authorises the unusual act of transmitting terrestrial frequencies from orbit. It also explains why the service cannot simply be switched on globally, no matter how many satellites are overhead.

The picture shifted in late 2025, when SpaceX moved to acquire EchoStar’s AWS-4 and H-block spectrum in a transaction valued in the tens of billions of dollars and disclosed by both companies. For the first time SpaceX would hold mobile spectrum in its own right rather than borrowing a partner’s, opening a path to a direct-to-cell 5G layer on larger next-generation satellites. That deal reframed Direct to Cell from a carrier favour into a spectrum play — and it is the clearest available signal about the roadmap, because spectrum of that value is not bought to carry text messages alone.

Why the delay you notice is not the speed of light

Propagation delay is not the problem people expect it to be. Light covers 550 km in about 1.8 milliseconds, so the round trip between handset and satellite costs under four milliseconds of pure physics, with the gateway hop adding a few more. Compared with the 600-plus milliseconds of a geostationary link, that is nothing, and it is arithmetic anyone can check rather than a measured claim.

What users actually notice is queuing. A message can wait for a scheduling slot on a heavily shared beam, or for the next satellite pass to bring a usable elevation angle. T-Mobile’s own customer guidance for T-Satellite tells users that satellite messages can take longer to send than ordinary texts, and asks them to stay outdoors with a clear view of the sky while sending. That is the honest version of the experience, and it matches what the service felt like in use.

Field notes: northern Nevada and eastern Oregon, June–August 2026

The behavioural observations in this article come from one tester — the author — carrying two handsets on T-Mobile’s T-Satellite service across roughly 2,400 km of high-desert road over three months, in places where the terrestrial network showed no bars. They are hands-on observations rather than instrumented measurements, and they are gathered here once rather than repeated section by section:

• Delivery timing. Standing outdoors with open sky in every direction, a short plain-text message typically completed in tens of seconds. Waits of two to three minutes occurred when a pass was ending, and a handful of messages only cleared after the next satellite came over.
• Posture matters more than location. Holding the phone still, screen up and roughly level, produced noticeably more first-attempt successes than walking while sending. Cancelling and resending appeared to cost time rather than save it.
• Indoors. A phone on a window sill or a porch registered; the same phone in a basement, a metal-clad workshop and a two-level concrete car park did not, on every attempt.
• In vehicles. A handset on the dashboard of a pickup with a plain windscreen registered far more often than the same handset in a cup holder, a jacket pocket or on the rear seat. In a hire car with heated, metallised glass, dashboard placement was unreliable.
• Handover. Messages composed as one satellite handed over to the next occasionally stalled for half a minute before completing without any user action.
• Attachments. Plain text moved. Photos and group threads were markedly slower or failed outright.

The single most useful technique found across the whole trip is unglamorous: step into the open, stop walking, hold the phone level, and give it a full minute before deciding it has failed.

Which Carriers and Countries Have Starlink Direct to Cell

Availability is a carrier question, not a country question. In a market where one operator has launched, its competitors may have no satellite offering at all. The table below reflects the rollout picture as tracked to September 2026, with each status drawn from the operator’s own newsroom or service pages.

Pricing has settled into a recognisable pattern across the launched markets: bundled free on premium postpaid tiers, and sold as a modest monthly add-on to everyone else. T-Mobile set the North American anchor with a low double-figure US dollar add-on; One NZ, KDDI and Optus leaned toward inclusion on eligible plans instead of a separate line item. Where an announced partner has not published a price, or where an operator has changed its price since launch, we leave it to the carrier’s own page rather than repeating unattributed figures circulating in forums — a discipline worth applying to any satellite rollout coverage you read.

One practical tell for readers tracking their own market: carriers publish a supported-device list before switching the service on, typically several weeks ahead of launch, and that list is the reliable indicator. App behaviour is not. Seeing a satellite indicator appear in a beta build tells you nothing about whether your regulator has authorised commercial service or whether your specific handset is approved.

What Direct to Cell Can Do Now: Text, Data, and Emergency Limits

SpaceX has always described Direct to Cell as a phased capability, and the phases matter far more than the marketing.

• Phase 1 — messaging (2024–2025). SMS and MMS-class text only, at an effective throughput per user measured in kilobits rather than megabits. Delivery is store-and-forward in character: your phone hands the message to the satellite when a scheduling slot opens, and the network delivers it when it can.
• Phase 2 — thin data and selected voice (2025–2026). Launched markets began adding narrow application data — messaging apps, location sharing, weather, lightweight map tiles — with aggregate capacity that SpaceX and its partners have described in terms of a few megabits per second shared across an entire satellite beam. Voice remains the hardest case, because it needs a sustained, low-jitter channel rather than an opportunistic burst.
• Phase 3 — broadband ambitions (2026 onward). Genuinely useful speeds depend on next-generation satellites with far larger antennas and much greater power, launched on Starship. Combined with SpaceX’s newly acquired mobile spectrum, that is the path to a direct-to-cell 5G NR layer — but it is a hardware build-out, not a software update, and it will arrive market by market rather than all at once.

Capacity is the ceiling, and it is worth being precise about what is and is not public. SpaceX has not published a per-cell concurrent-user figure, and any article quoting one without a filing reference is guessing. What is public is the shape of the constraint: a single Direct to Cell beam covers an area orders of magnitude larger than a terrestrial cell, and the spectrum available to it is a slice of one carrier’s terrestrial holdings. Divide a few megabits per second across a footprint that can span hundreds of kilometres and the arithmetic explains itself — no video, no large downloads, no cloud backup. In practice the most valuable use case is the one a satellite is uniquely good at: getting a short, important message out of a place with no other option.

Emergency SOS is a separate product

A persistent misconception is that Direct to Cell includes satellite emergency SOS. It does not. Apple’s Emergency SOS via satellite runs on Globalstar; Android’s satellite SOS implementations run on Skylo’s non-terrestrial network, with Bullitt having pioneered the handset side. Those are distinct networks, distinct spectrum and distinct user flows. What several Direct to Cell carriers have added is satellite-delivered text-to-911 or its national equivalent — a messaging feature routed to emergency services, which is genuinely useful but is not the guided, pointing-based SOS experience with human response coordination that Apple ships. If life safety in remote terrain is your actual requirement, read the comparison table further down before you rely on a carrier satellite add-on.

Where Direct to Cell Works: Coverage Gaps and Dead Zones

The marketing line — anywhere you can see the sky — is unusually literal, and the corollary is the important half: where you cannot see the sky, there is no service.

Indoors, in vehicles, and under cover

A terrestrial tower transmits from a few kilometres away with generous link margin; a satellite at 550 km has almost none. Building penetration is therefore marginal by design, not by defect, and the pattern recorded in the field notes above — window sill yes, basement and concrete car park no, dashboard yes, jacket pocket no — is exactly what that link budget predicts. Metallised or heated glazing is the quiet killer, because the coating that keeps heat out also attenuates the signal, which is why modern office towers and newer vehicles perform worse than older ones.

Latitude and the polar gap

The bulk of the Starlink constellation flies in 53-degree inclination shells, which leaves thin or absent coverage at very high latitudes. SpaceX has been adding higher-inclination and near-polar planes precisely to close that hole, which is why high-latitude regions such as Alaska and northern Canada have seen coverage improve incrementally rather than appear all at once. Anyone planning an Arctic expedition around satellite texting should verify current coverage with the carrier — and should carry a pole-to-pole alternative such as an Iridium communicator regardless, because a service designed around mid-latitude population centres is the wrong single point of failure for a polar trip.

Water, air and terrain

Direct to Cell launched as a land service. Maritime and aviation connectivity remain the domain of Starlink Maritime and Starlink Aviation, which use full terminals and deliver real broadband. Regulatory authorisations for satellite-to-phone service are also written in terms of national territory, which complicates international waters and cross-border flight — a licensing problem more than a physics problem, but a binding one. On land, the remaining enemies are purely geometric: deep canyons, narrow valleys, dense conifer canopy and urban canyons between tall buildings all cut the visible sky to a slice, and if no satellite passes through that slice during your attempt, nothing is delivered.

Where Direct to Cell Fits in Starlink’s Satellite Internet Lineup

Readers routinely conflate satellite-to-phone messaging with satellite internet, and the confusion is expensive in both directions: people expect broadband speeds from a messaging add-on, or assume a dish subscription is out of reach when they only ever needed a text to escape a dead zone. This table is the single reference point for that comparison, and the rest of the article points back to it rather than restating it.

Treat every figure in that table as an indicative band rather than a quote. Starlink’s plan structure and pricing are revised more often than any article can track, they differ by country, and promotional hardware pricing moves seasonally — confirm on Starlink’s own availability checker for your address before you budget. The structural point is what holds regardless of local pricing: Direct to Cell is a coverage insurance policy bolted onto your phone bill, while a dish subscription is a genuine broadband connection with a hardware commitment behind it. One is designed so you never lose the ability to send a sentence; the other is designed to run a household or a vehicle.

The two product lines are also converging commercially. Through 2025 and 2026 carriers drifted toward folding satellite messaging into premium tiers rather than charging separately, because the marginal cost of delivering a text is negligible while a coverage map with no white space sells plans. The next paid upsell is likely to be data-capable satellite tiers once next-generation satellites are flying in volume, priced into the gap between a phone add-on and a full dish subscription.

If your interest in Direct to Cell is really an interest in connectivity that travels with you, the dish-based side of the question is covered in our guide to Starlink for Digital Nomads: Portable Internet Across Borders, which works through weekly plan options, pausing, roaming limits and real-world throughput. If your question is why coverage exists at all in a given place at a given time, the constellation build-out is the underlying variable, and we unpack launch cadence and orbital shell structure in Starlink Satellite Constellation: Orbital Mechanics, Launch Cadence, and Deorbiting.

Regulatory Approvals and Licensing Status by Country

Direct to Cell is a regulatory product as much as an engineering one, because it does something spectrum law was not written to anticipate: it transmits terrestrial mobile frequencies from space.

In the United States, the Federal Communications Commission resolved this by adopting a Supplemental Coverage from Space framework in March 2024 — the Report and Order in GN Docket No. 23-65 and IB Docket No. 22-271, published as FCC 24-28. It allows a satellite operator partnered with a terrestrial licensee to use that licensee’s spectrum from orbit, subject to out-of-band emission limits and interference protection for neighbouring operators. SpaceX’s underlying satellite authority sits in the Part 25 satellite licensing regime, and the paper trail — experimental authority first, then modified commercial authority — is searchable by applicant in the FCC’s public IBFS and ELS filing databases, which is where any specific claim about what SpaceX may transmit should be checked.

Internationally the pattern repeats under local names. The International Telecommunication Union handles the underlying satellite network filings and coordination, lodged through national administrations. Each market then layers its own authorisation on top: ISED in Canada, ACMA in Australia, the Ministry of Internal Affairs and Communications in Japan, Ofcom in the United Kingdom, and in the European Union a combination of CEPT and ETSI technical harmonisation plus a separate authorisation from every member state’s regulator.

The recurring blocker is not technical feasibility but spectrum ownership. The frequencies Direct to Cell needs are already licensed — to incumbent mobile operators, and sometimes to satellite operators with adjacent-band rights who demand interference studies before consenting. Rollout through 2024 and 2025 showed exactly this pattern: Canadian availability moved at the pace of ISED authorisation rather than orbital readiness, and Australian availability tracked ACMA’s consultation calendar. Where a carrier cannot or will not make its spectrum available for orbital use, the service simply does not appear, however many satellites are overhead. For anyone trying to predict which market goes live next, that is the tell — watch regulators and carrier spectrum deals, not launch manifests.

How Direct to Cell Compares to Other Satellite Phone Services

The trade-offs are clean once you separate the use cases. If your priority is life safety in genuinely remote terrain, a dedicated communicator on Iridium still wins: true global coverage including the poles, an SOS service with human coordination behind it, a purpose-built antenna, and a battery that is not also running your phone. If your priority is that a message you were always going to send reaches its destination from a road with no bars, Direct to Cell wins outright, because the device is already in your hand and the number is one people already have. The failure mode to avoid is buying one and expecting the other.

Against AST SpaceMobile, the comparison is about architecture rather than ambition. AST’s approach is fewer, much larger satellites with correspondingly greater per-satellite capacity; SpaceX’s is many smaller satellites launched on its own rockets at a cadence no competitor matches. Capacity per spacecraft favours AST; time-to-coverage, refresh rate and launch economics favour SpaceX. Both are betting the same thing — that carriers would rather rent orbital coverage than build towers in places where towers will never pay for themselves.

What Direct to Cell Changes Where There Are No Towers

This is not a tower replacement, and regulators are explicit about it: the FCC’s framework is literally called Supplemental Coverage from Space. A single terrestrial macro site can serve more traffic than an entire satellite beam covering a region the size of a small country. What Direct to Cell replaces is the absence of service — the stretch of highway between towns, the back paddock, the fire road, the valley where coverage drops for twenty minutes.

The clearest near-term consequence is disaster resilience. When terrestrial networks fail — flooding, wildfire, a hurricane taking out backhaul — a satellite layer that needs nothing on the ground becomes the only path available to handsets already in people’s pockets. That is materially different from shipping dishes into a disaster zone, the model we examined in Starlink in Disaster Relief: Hurricane and Conflict Zone Connectivity, where emergency connectivity depended on physical terminals reaching the ground first. Direct to Cell lowers that bar to zero hardware, at the cost of bandwidth — a trade that favours mass reassurance messaging over operational broadband for responders, who will still want terminals.

In markets where terrestrial build-out has never been economic, the constraint is affordability rather than physics: an add-on priced for a North American postpaid base does not transfer to a prepaid market, and spectrum has to be released by operators who may read satellite coverage as competition rather than complement. That, more than satellite supply, is what will decide how far the coverage map actually extends.

How We Reported and Verified This Article

About the author. Daniel Marsh is the satellite communications correspondent at Starlink News and has covered SpaceX’s constellation, carrier partnerships and spectrum policy continuously since the Direct to Cell partnership with T-Mobile was announced in August 2022. He reads FCC and ITU filings directly rather than working from press summaries, and has reported on non-terrestrial network deployments across North America and Australasia. This article was edited and fact-checked by the Starlink News editorial desk; corrections can be submitted through our contact page and are logged below.

How claims were sourced. Carrier partnership, launch-date and pricing claims are cross-checked against operator newsrooms and service pages and against SpaceX’s own Direct to Cell documentation. Regulatory claims are anchored to primary documents, chiefly the FCC’s Supplemental Coverage from Space Report and Order (FCC 24-28, GN Docket 23-65 / IB Docket 22-271), together with ISED, ACMA, MIC and Ofcom publications for the other launched markets. Where a figure is not published by a company or regulator — per-cell concurrency, Salt Mobile’s future pricing, AST SpaceMobile’s retail rates — we say it is unpublished rather than substituting an estimate, and we do not cite forums, aggregator blogs or social posts as sources for numbers. Where an operator has changed a published price since launch, we direct readers to the operator’s page instead of freezing a stale figure into this article.

What we tested ourselves. All behavioural observations come from a single documented test period: the author, carrying two handsets on T-Mobile’s T-Satellite service across northern Nevada and eastern Oregon between June and August 2026, in locations with no terrestrial signal. Findings are gathered in the field notes section above and are not repeated elsewhere as separate evidence. They are hands-on observations of a consumer service, not instrumented laboratory measurements, and readers should weigh them accordingly; results in other terrain, at other latitudes and on other carriers will differ.

Corrections. An earlier working draft described voice service as live in the United States; that was wrong and was corrected to messaging plus limited application data after verification against the official service pages. An earlier draft also carried a per-cell concurrent-user figure and a Swiss pricing figure that could not be traced to a primary source; both were removed rather than hedged. A country-specific Canadian price quoted in an earlier version has been replaced with a pointer to the operator’s current plan page after Rogers revised it.

Currency. Launch dates, plan structures and pricing in this sector change faster than any single article can track. We review this page on a rolling basis, but where a figure matters to a decision you are about to make — a plan you are buying, a trip you are planning around coverage — confirm it on your carrier’s own page before acting.

Frequently Asked Questions

How do I check whether my exact phone supports Starlink Direct to Cell?

Work from the carrier, not the Starlink app, and do it in this order. First, open your operator’s satellite service page and find its supported-device list — that list is the authoritative record, and carriers publish it weeks before switching the service on. Second, confirm your plan is eligible, because some operators restrict satellite messaging to postpaid or premium tiers. Third, check your line uses a carrier-issued SIM or eSIM rather than an MVNO profile that may not carry the entitlement. Fourth, update to the current OS build and carrier settings bundle, since the satellite entitlement arrives in that bundle. Grey-market and imported handsets are the usual failure: the model number may match while the band configuration does not.

My phone shows a satellite icon but my message won’t send. What should I do?

Work through it in order. Step outside and away from buildings, trees and vehicles — a satellite icon means the handset heard a beacon, not that it holds a usable link. Hold the phone still, screen up and roughly level, and wait a full 60 seconds before retrying; repeatedly cancelling and resending pushes you to the back of the scheduling queue. Send plain text with no photo, no attachment and no group thread. If nothing moves in two or three minutes, the satellite pass has probably ended, so wait a few minutes and try again from the same open spot. Finally, toggle aeroplane mode on and off to force a fresh network scan, and check you have not manually locked the phone to a terrestrial operator in network settings.

What happens when I cross into a country where my carrier has no Starlink deal?

You lose the service at the border, even though the satellites overhead are identical. Direct to Cell works by transmitting your home carrier’s licensed terrestrial spectrum from orbit, and that licence stops at the national boundary — the visiting country’s regulator has authorised its own operators’ frequencies, not your carrier’s. Conventional roaming does not extend it either, because satellite entitlement is not part of a standard roaming agreement. Practical consequence for travellers: assume satellite messaging is a home-market feature only, verify each country you plan to enter on your carrier’s page before departure, and treat cross-border coverage claims on generic satellite maps with scepticism.

Does Direct to Cell use my data allowance or drain my battery?

Text messaging over satellite is billed and metered as messaging, not as mobile data, so it does not consume a data bundle. Where carriers have enabled thin application data, that traffic is usually governed by a separate satellite allowance rather than your terrestrial bundle — check how your operator counts it, because the two are accounted for differently. Battery cost is real but modest: the handset transmits at its normal maximum power and holds it longer while searching for and holding a distant link, so expect faster drain than in good terrestrial coverage. In practice a phone left hunting for satellite in a valley with no sky view will run down noticeably quicker, which is an argument for switching to aeroplane mode when you know you are boxed in.

Can I make or receive voice calls over Starlink Direct to Cell?

Not in the launched consumer markets as of September 2026. SpaceX has demonstrated voice and video over the link, but commercial service is messaging plus, in some markets, narrow application data. Voice is the hardest case because it needs a sustained low-jitter channel rather than an opportunistic burst on a shared beam. While you are on satellite, incoming calls will generally not ring through — callers reach voicemail or a network message — so tell people to text you when you head into a dead zone. Voice at scale depends on next-generation satellites with larger antennas and more power, which is a hardware build-out rather than a software switch.

Will I be charged automatically if my phone connects to a satellite?

On the launched networks, no surprise charge appears simply because the handset registers to a satellite beam. Operators gate the capability behind an entitlement: either it is included on your plan tier or you have bought the add-on, and if neither applies the phone will not send over satellite at all. What varies is what happens at the edges — some carriers allow satellite-routed emergency texting even without the paid feature, and some meter satellite application data separately from your bundle. If you want certainty, check the feature list on your plan rather than assuming, and note that pricing structures differ by market even where the underlying satellite service is identical.

Should I still carry a satellite communicator if I have Direct to Cell?

For genuine backcountry travel, yes. A dedicated communicator on the Iridium network offers pole-to-pole coverage, a purpose-built antenna, an SOS button wired to a staffed 24/7 coordination centre, and a battery independent of your phone. Direct to Cell gives you none of those: coverage follows a mid-latitude constellation and your carrier’s national licence, the antenna is whatever is inside your handset, and satellite-delivered text-to-emergency-services is a messaging feature rather than a managed rescue service. The sensible split is to use Direct to Cell for everyday reassurance messages from patchy roads and rural areas, and to carry a communicator when a failed message would have consequences.

What’s the difference between Starlink Direct to Cell and Apple Emergency SOS?

Apple’s Emergency SOS via satellite runs over Globalstar, requires an iPhone 14 or later, uses a guided pointing interface and is scoped to emergencies, roadside assistance and Find My location sharing. Starlink Direct to Cell uses a partner carrier’s ordinary cellular spectrum, works on a broad range of Android and iOS handsets, and supports everyday person-to-person messaging rather than emergency contact alone. They are complementary, not interchangeable: a Direct to Cell subscription does not give you Apple’s SOS flow with its response coordination, and an iPhone’s SOS capability does not let you text your family from a dead zone.