O2 Satellite

1. What O2 Satellite Actually Is – and What It Isn’t

On 26 February 2026, Virgin Media O2 switched on O2 Satellite – a direct-to-device service built on SpaceX’s Starlink Direct-to-Cell (DtC) network. It made O2 the first mobile operator in the UK and Europe to offer satellite-powered data services. The government called it a “major achievement.” The press called it a revolution. Both were right – and both slightly missed the point.

The headline numbers are real. O2’s UK landmass coverage jumped from 89% to 95%. That’s an uplift described as roughly two-thirds the size of Wales. For consumers hiking in the Brecon Beacons or sailing off the Pembrokeshire coast, this is genuinely transformative. You don’t need a special terminal. You don’t need a satellite phone. If you have a compatible handset, the network finds the satellite automatically when terrestrial signal drops out.

For M2M and IoT professionals, the launch is highly significant – but not for the reasons the press coverage suggests. The real significance is what it signals about the direction of travel, not what you can deploy today. To understand why, you need to understand what the technology actually does.

The service launched with immediate restrictions. Device support at launch is limited to the Samsung Galaxy S25 range – the S25, S25+, S25 Ultra, and S25 Edge. App support covers a defined set including WhatsApp, Messenger, Google Maps, AccuWeather, BBC Weather, AllTrails, and a handful of others. Standard cellular voice calls are not supported at launch. WhatsApp calls work. SMS and limited data work. That’s your starting point.

There is another restriction you won’t find prominently in the press releases. O2 Satellite currently operates only up to the 58th parallel north. That’s roughly a line running through central Scotland – somewhere around Dundee and Stirling. The Highlands, the Northern Isles, and the far north of Scotland remain outside the service area for now. O2 has said this will improve as next-generation satellites are deployed. SpaceX plans to launch its Gen 2 DtC satellites in late 2027, which are projected to deliver 5G connectivity and peak speeds of 150 Mbps per user.

2. How Starlink Direct-to-Cell Works at a Technical Level

The term “cell tower in space” gets used a lot. It’s a useful shorthand. But what it actually means technically is worth unpacking, because the engineering is genuinely novel – and understanding it explains both the capabilities and the constraints.

The satellite as a moving base station

Standard Starlink uses Ku-band and Ka-band spectrum for fixed broadband. DtC is entirely different. The DtC-capable satellites – currently around 650 of Starlink’s constellation of approximately 9,800 – carry large phased-array antennas specifically designed to broadcast in licensed terrestrial mobile spectrum. In the UK, early monitoring of the live service by ISPreview.co.uk identified the service operating on Band 3 (1800 MHz). O2 holds 1800 MHz spectrum, and this is what the satellites are transmitting.

This is not a trivial choice. 1800 MHz is widely supported across the IoT module ecosystem – most LTE-capable devices support Band 3. The signal also has better indoor penetration characteristics than higher frequency bands, though significant attenuation through walls remains a real-world limitation.

Each DtC satellite is travelling at approximately 17,000 mph at an altitude of around 550 km. To a device on the ground, each satellite is a base station that appears, passes overhead, and hands off. To the LTE device, it should look like a standard cell – an eNodeB. Getting to that point requires solving two serious physics problems.

Doppler compensation

When a transmitter is moving relative to a receiver, the frequency of the signal shifts. At orbital velocities, this Doppler shift is substantial – far beyond what standard LTE receivers are designed to handle. Starlink’s solution is to pre-correct the signal at the satellite. The satellite’s software-defined radio calculates the Doppler shift for every device in its coverage footprint and adjusts the transmitted frequency accordingly. The device on the ground receives a signal that appears to be coming from a stationary source.

Timing advance and propagation delay

Standard LTE base stations are terrestrial. The round-trip propagation delay is measured in fractions of a millisecond. At 550 km altitude, even at the speed of light, you’re looking at round-trip delays of 30-80 ms depending on the satellite’s angle of elevation. LTE uses Timing Advance (TA) to synchronise uplink transmissions. For satellite links, this requires extended TA values well beyond normal LTE parameters – another problem solved in the satellite’s processing stack.

The end result is that, to a compatible device, the satellite presents itself as a standard O2 cell, identifiable on MCC 234, MNC 02. The handshake happens automatically. No app, no manual configuration, no user action required.

Backhaul

The satellite doesn’t just transmit to devices – it needs to route that data somewhere. Starlink uses inter-satellite laser links to pass data between satellites, eventually routing it down to ground stations that connect into O2’s core network. This is the same laser mesh that makes Starlink’s fixed broadband competitive on latency. For the DtC service, it means the backhaul path is largely independent of terrestrial infrastructure – which is exactly the point.

3. The Awkward IoT Reality Check

Here is the thing nobody in the consumer press has said clearly: O2 Satellite currently has no IoT or M2M capability whatsoever.

It runs on Samsung Galaxy S25 handsets only. It does not support IoT modules. It does not support industrial routers. It does not support any M2M device category. The service that launched in February 2026 is, in its current form, a consumer product for people who want WhatsApp access on a hill in Wales.

That is not a criticism. It’s the correct place to start. Every network technology goes through this phase. The M2M implications come later – but you need to understand what “later” means before you plan around it.

Why IoT modules can’t use it yet

Starlink DtC works by making the satellite appear as a standard LTE cell to a compatible device. “Compatible” is the operative word. The handset-side requirements include specific modem firmware, specific baseband chipset support for extended Timing Advance values, and in some cases OS-level integration for the automatic handoff behaviour. IoT modules – the embedded cellular modems used in routers, trackers, meters, and sensors – are built to different specifications. They do not have the same firmware capabilities. They have not been validated for DtC service.

This will change. Qualcomm and MediaTek both have roadmaps for modem firmware that will support DtC connectivity. The 3GPP has been standardising this at a protocol level. But “on the roadmap” is not “available today.” Don’t conflate the two.

The satellite congestion question

Each DtC satellite footprint covers a large geographic area. Unlike a terrestrial base station serving a specific cell, a satellite passing overhead is simultaneously serving every compatible device within a radius of hundreds of kilometres. Capacity is shared across all of those devices. In a remote agricultural setting, that might be fine. In an area where thousands of handsets are suddenly falling back to satellite during a terrestrial outage, congestion management becomes critical. MNOs will prioritise traffic based on QoS Class Identifier (QCI). Consumer messaging traffic and emergency calls will sit at higher priority than unclassified M2M data. This matters for any future IoT use case on satellite.

What the roadmap actually looks like

SpaceX’s Gen 2 DtC satellites, planned for late 2027, will significantly expand capability. The stated target is 5G connectivity with peak data speeds of 150 Mbps per user. That’s a different service category entirely. Alongside device support expansion – iPhone support is conspicuously absent at launch, and Apple’s entry would massively increase the addressable market – the industrial and IoT ecosystem should begin to see compatible module hardware emerge from around 2027 onwards. Plan for 2028-2029 as the realistic window for production-grade IoT deployment on DtC infrastructure in the UK.

4. The Two NTN Paths – and Why They Are Completely Different Technologies

This is where the coverage gets muddled. Almost every article about O2 Satellite conflates two genuinely separate technologies under the umbrella of “satellite IoT.” They share a general direction but differ in almost every technical dimension that matters for M2M deployment decisions.

The critical distinction: Path A works with your existing Cat-1 or Cat-4 module once firmware support arrives. Path B requires you to buy new hardware specifically designed for IoT NTN, with GNSS capability built in. They address different use cases and different hardware generations.

The practical implication: if you are designing a new IoT deployment today and satellite coverage is on your requirements list, you need to choose your path. A sensor that transmits a small payload every few hours – a water level sensor, a soil moisture probe, a livestock tracker – is a candidate for IoT NTN hardware today. A cellular router providing connectivity for a remote SCADA terminal needs to wait for DtC module support, which is a 2028 conversation at the earliest for reliable production deployment.

5. AST SpaceMobile and Vodafone SatCo – The Other Play

Starlink is not the only game in town for direct-to-device satellite. AST SpaceMobile, backed by Vodafone, AT&T, Verizon, Google, and others, is building a competing constellation on a fundamentally different architecture.

Where Starlink uses a large number of smaller satellites, AST SpaceMobile is using a smaller number of very large ones. Its Block 1 BlueBird satellites each carry a phased-array antenna spanning 64 square metres. The upcoming Block 2 generation will carry antennas up to 2,400 square feet – roughly 223 square metres. The antenna is so large it caused controversy with astronomers due to its brightness. But size is the point: a bigger antenna means more gain, which means the ability to communicate with standard unmodified handsets without the base station needing to be the one doing all the heavy lifting.

The claimed performance numbers are different too. AST SpaceMobile has demonstrated over 20 Mbps download speeds to unmodified phones on a 5 MHz channel. Its Block 2 satellites are designed to deliver peak data rates up to 120 Mbps. That’s genuine broadband from space – not the low-bandwidth messaging service that DtC launched with.

Vodafone SatCo

In March 2025, Vodafone and AST SpaceMobile formalised a joint venture called SatCo, headquartered in Luxembourg. SatCo’s stated purpose is to offer wholesale direct-to-device mobile broadband to all European MNOs. Operators in 21 EU member states have expressed interest. Commercial launch is planned for 2026, though precise timelines have remained broad. AST SpaceMobile needs 45-60 satellites in orbit for meaningful commercial coverage, and it is aiming for that number by end of 2026.

For UK deployments, this means Vodafone UK is likely to follow O2 with its own satellite service – but on AST SpaceMobile’s technology rather than Starlink’s. The two services will have different capability profiles. If AST’s broadband numbers hold up at scale, the Vodafone service could represent a meaningfully higher-throughput option than O2 Satellite’s current messaging-focused capability.

6. 3GPP IoT NTN – The Standardised Route for Real M2M

Away from the consumer headlines, a quieter and arguably more important development has been unfolding. 3GPP Release 17, finalised in 2022, introduced Non-Terrestrial Network (NTN) support for both NB-IoT and LTE-M (eMTC). This is the standardised, interoperable path for massive IoT connectivity via satellite. It is not a workaround. It is not proprietary. It is the long-term foundation.

What 3GPP IoT NTN actually standardises

The challenges of connecting IoT devices to satellites are similar to the handset case – Doppler shift, extended propagation delay, large timing advance requirements – but with additional constraints. IoT devices run on batteries for years. They transmit tiny payloads infrequently. They do not have the processing power of a smartphone. The solutions have to be extraordinarily power-efficient.

3GPP’s approach for IoT NTN follows the same principles as NR-NTN (the 5G version) but adapted for NB-IoT and LTE-M. The key requirement is that IoT NTN devices must have GNSS capability. This is not optional. The device uses its own position data and the known satellite ephemeris to pre-calculate the Doppler shift and timing advance, and applies uplink pre-compensation autonomously. It solves the synchronisation problem without needing the network to do all the work. The implication is that every IoT NTN module will include GNSS – adding modest cost and current draw, but enabling operation that is simply impossible otherwise.

Discontinuous coverage – not a bug

One of the most interesting aspects of 3GPP’s IoT NTN work is the explicit recognition that discontinuous coverage is acceptable for many use cases. A sensor on a water management installation in the Scottish Highlands doesn’t need a permanent connection. It needs to successfully transmit a reading once every hour or once every four hours. A sparse LEO constellation might only have a satellite overhead for 10 minutes in every two hours. For this use case, that’s fine – and the standard has been designed to accommodate it. Store-and-forward operation, where the device buffers data and transmits when a satellite is available, is explicitly supported.

Who is building IoT NTN hardware today

Several chipset and module manufacturers are already in the market. Nordic Semiconductor has been working on NTN-capable variants of its nRF91 series. Sony’s Altair semiconductor division developed the ALT1350 chipset, one of the first to support IoT NTN. On the network side, operators including Skylo (which works with multiple MNOs globally), Sateliot (focused on NB-IoT LEO), and OQ Technology are already running commercial or near-commercial services using standardised IoT NTN technology.

This is the realistic hardware path for agricultural sensors, remote utilities monitoring, maritime tracking, and cold-chain logistics in the 2025-2028 window. Not DtC. Not O2 Satellite. Dedicated IoT NTN hardware designed specifically for this use case.

7. The Consumer SIM Threat – and Why It Falls Apart Under Scrutiny

Here’s a question that is already being asked in procurement meetings. If a standard O2 consumer SIM now has satellite coverage as part of the service, why pay a premium for an IoT-specific SIM? Add an edge platform like IoTInix or a managed connectivity layer, and you’ve got the IoT bits covered. It looks compelling on paper. It fails in production.

This is not a hypothetical concern – it’s happening. Consumer SIMs are being classified as “business SIMs,” dropped into industrial routers, and deployed with edge platforms doing the heavy lifting. Short-term this works. Medium-term it generates support tickets. Long-term it generates failures. Here’s why.

Kill switch 1: Fair usage and permanent roaming policies

Consumer SIM contracts are governed by fair usage policies. These policies exist to prevent commercial exploitation of consumer-priced products. One of the most relevant clauses covers permanent roaming: if a device connects primarily through a partner network or a non-terrestrial pathway for an extended period – typically 60 to 90 days – the MNO has the right to throttle or terminate the line.

For a device deployed in a remote location with poor terrestrial coverage, this is precisely the scenario you will encounter. The SIM spends most of its time on satellite or on a partner network cell. After 90 days, the line gets flagged. You get no notification. The device goes silent. You discover it during a manual audit, not because any alarm fired. Industrial IoT SIMs have no permanent roaming thresholds. They are designed for exactly this use case.

Kill switch 2: CGNAT and the static IP problem

Consumer SIMs use Carrier-Grade NAT (CGNAT). The device shares an IP address with potentially thousands of other devices behind the same NAT gateway. You cannot initiate a connection inbound to the device. You cannot dial in to check status, push configuration, or retrieve data without the device initiating the connection first. For many IoT applications this is workable. For anything requiring device management, remote diagnostics, or VPN-based secure backhaul – as used with any competent edge platform – it is a fundamental blocker.

Industrial IoT SIMs can be provisioned with static private IPs on a dedicated APN. The device gets an address that belongs to it and only it. VPN tunnels terminate correctly. Device management platforms can reach the SIM directly. This is not a luxury feature – it is a basic operational requirement for professionally managed IoT deployments.

Kill switch 3: QoS class identifier and satellite congestion

When a satellite footprint is congested – a terrestrial outage has pushed thousands of devices onto the satellite at once – the MNO applies traffic prioritisation. Emergency calls sit at QCI 1. Consumer messaging gets its own priority. Background data from a cheap consumer SIM sits at the bottom. Your MQTT heartbeat from a remote monitoring installation gets deprioritised and dropped. You don’t see this in testing. It appears in production during exactly the scenario where satellite connectivity matters most.

Industrial IoT SIMs carry a higher QCI assignment. The network treats their traffic differently. This is negotiated between the IoT SIM provider and the MNO at a wholesale level and baked into the SIM profile.

Kill switch 4: Hardware form factor and lifecycle

Consumer SIMs come in standard plastic plug-in form factors – 2FF, 3FF, 4FF. These are fine for offices and homes. In industrial environments – plant vibration, temperature cycling between -40°C and +85°C, moisture ingress – the connector oxidises, the card unseats, contact resistance increases, and eventually the connection fails intermittently. Intermittent SIM failures are among the hardest faults to diagnose remotely.

MFF2 (Machine Form Factor 2) embedded SIMs – soldered directly to the PCB – are the industrial standard. They are vacuum-sealed, vibration-tested, and rated for operational lifetimes of 10 years and beyond. No industrial IoT SIM provider sells you a plug-in consumer SIM for a 10-year deployment in a remote monitoring installation. The form factor is part of the product.

The last row is worth sitting with. Neither consumer SIM nor industrial IoT SIM can give you satellite connectivity on an IoT module today. The consumer SIM has the service in theory – on a phone you’re not deploying. The industrial IoT SIM doesn’t have it yet either, but comes with every other operational advantage. When satellite access does arrive for industrial hardware, it will come through the industrial SIM ecosystem first – because that’s where the commercial relationships and the QoS agreements sit.

8. The MVNO Capability Gap – Why IoT SIM Providers Can’t Offer This Yet

A legitimate frustration in the M2M SIM market is the pace at which new network capabilities reach the MVNO and IoT SIM provider ecosystem. 5G Standalone is a clear example. Many IoT SIM providers cannot offer 5GSA today, years after MNOs switched it on. The reasons are structural, not laziness. The same structural barriers apply – more acutely – to satellite access.

The Light MVNO problem

Most IoT SIM providers in the UK market operate as Light MVNOs. They own SIM cards and commercial relationships with MNOs, but they do not own their own Home Location Register (HLR) or Home Subscriber Server (HSS). They sit on top of the MNO’s core network infrastructure. When the MNO upgrades its core, the Light MVNO inherits whatever the MNO exposes through its wholesale interface – and nothing more.

5G Standalone runs on a Service-Based Architecture (SBA), using HTTP/2 signalling between network functions. The equivalent of the HLR in a 5G core is the Unified Data Management (UDM) function, and inter-operator interfaces use a Security Edge Protection Proxy (SEPP). Most Light MVNOs are still running on 4G Diameter signalling. They have not invested in the 5G core infrastructure needed to peer correctly with an MNO’s 5G network functions. Until they do, they cannot offer 5GSA services – regardless of whether the underlying network supports them.

The satellite access problem is worse

Satellite connectivity via Starlink DtC is not just a 5G core issue – it’s a spectrum access issue. O2 Satellite uses O2’s licensed spectrum, broadcast from space by Starlink’s satellites. That spectrum is O2’s. The right to use it for DtC service is enshrined in a specific licence variation granted by Ofcom in February 2026. O2 chose to offer this to consumers and enterprise customers directly. They have not announced any wholesale arrangements to allow MVNOs or IoT SIM providers to resell satellite access.

This is not an oversight. It is deliberate. MNOs view satellite coverage as a premium differentiation factor. It is the headline feature they use to justify enterprise contract pricing. Offering it through wholesale channels – at MVNO margins – would undermine that positioning. The commercial logic for keeping satellite access proprietary is strong, at least in the short term.

The eUICC workaround path

Some IoT SIM providers are exploring a route that bypasses the MNO bottleneck entirely: eUICC (embedded SIM) with multi-IMSI profiles. The principle is that the SIM can hold multiple operator profiles. When a device moves into a zone with poor terrestrial coverage, the SIM management platform detects this and pushes a profile from a different operator – potentially one with satellite access – onto the SIM. The device effectively changes networks without changing hardware.

This is technically feasible under the 3GPP eUICC specification. The practical barriers are commercial: you need a relationship with a network that actually has satellite IoT access to sell, and those relationships don’t yet exist at scale in the UK market. Skylo, for instance, has operator relationships and satellite IoT NTN capability. An IoT SIM provider with an eUICC platform and a Skylo wholesale agreement could theoretically offer satellite fallback today – but for NB-IoT/LTE-M NTN payloads, not DtC broadband.

What Full MVNO status actually buys you

A Full MVNO – one that owns its own HLR/HSS, its own packet core, and its own roaming agreements – has more leverage. It can negotiate directly with satellite operators and with MNOs for specific roaming agreements that include satellite access. It can implement its own QoS policies. It can offer static IP addressing on its own APN. This is why the leading industrial IoT SIM providers have been investing in Full MVNO infrastructure rather than remaining on Light MVNO wholesale agreements. The satellite era will reward those who made that investment.

9. The Roadmap – What Happens Between Now and 2030

Predicting the exact path of satellite IoT is somewhere between educated analysis and reading tea leaves. But the structural forces at work are clear enough to outline a credible trajectory.

The 5GSA parallel

The best analogy for how satellite access will reach the IoT SIM market is the 5G Non-Standalone to 5G Standalone transition. 5G NSA launched commercially in 2019. Here we are in 2026, and many IoT SIM providers still cannot offer 5GSA to their customers. The reasons are the same: core network investment, roaming agreement complexity, commercial positioning. Satellite access will follow the same slow institutional timetable. The technology will outrun the commercial infrastructure by several years.

This is not pessimism. It is pattern recognition. The appropriate response is to build IoT deployments on what is available now, with hardware that is upgradeable, on SIM platforms from providers who are investing in Full MVNO infrastructure and who have published clear satellite access roadmaps.

The multi-network SIM evolution

The most likely commercial model for satellite IoT access from the provider ecosystem is an extension of the multi-network roaming SIM concept that matured in the 2015-2022 period. A SIM that can access multiple terrestrial networks will, in time, extend that capability to satellite networks. The eUICC/GSMA SGP.02 (M2M eSIM) framework already supports remote profile management. Extending this to manage satellite network profiles is not a significant technical leap. The commercial leap – negotiating the agreements – is where the time and complexity lies.

The Scottish Highlands problem and what solves it

The 58th parallel limitation is a genuine near-term gap. The areas of the UK most likely to benefit from satellite connectivity for IoT – the Highlands, the islands, upland Wales, parts of Northumberland – are exactly the areas where coverage constraints exist. The Gen 2 Starlink satellites planned for 2027 are expected to extend coverage to higher latitudes. AST SpaceMobile’s constellation design has different orbital parameters and should have fewer high-latitude constraints. For deployments in Scotland’s most remote areas, 3GPP IoT NTN via a dedicated satellite provider like Skylo or Sateliot may be the more pragmatic near-term path.

Price compression

One theme that will run through the entire 2026-2030 period is price compression. Consumer satellite access at £3 per month signals the direction of travel. Specialist satellite IoT data – which has historically been priced at a premium – will face downward pressure as terrestrial satellite access becomes more mainstream. This is good for the market overall. It may squeeze margins for providers who have built businesses on high-cost satellite IoT tariffs without adding corresponding value.

10. Summary – What IoT Managers Should Actually Do Right Now

A lot of noise. Here is the signal.

If you are specifying a new IoT deployment today

Do not wait for DtC satellite module support before deploying. It is not ready. If your use case is a low-power sensor sending small payloads in an area with no terrestrial coverage, evaluate 3GPP IoT NTN hardware (NB-IoT satellite) from suppliers already in the market. If your use case requires a standard data pipe, deploy terrestrial cellular now with hardware and SIM platform that will accommodate satellite when the module ecosystem catches up – likely 2028 at the earliest for reliable production deployment.

If you are evaluating whether a consumer SIM will do the job

It won’t. Not for production. The failure modes are documented: permanent roaming thresholds, CGNAT, QoS deprioritisation, form factor. A consumer SIM on a bill marked “business” is a prototype with an invoice. Use an industrial IoT SIM from a provider with static IP, a private APN, no fair usage, and MFF2 hardware options. That’s the professional baseline before satellite enters the picture.

If you are evaluating IoT SIM providers

Ask two questions. First: are you a Light MVNO or a Full MVNO? Second: what is your satellite access roadmap, and which satellite operators do you have commercial relationships with? A provider running on a Light MVNO model with no eUICC capability and no satellite relationships is selling you connectivity that will hit a ceiling when satellite access becomes a customer expectation. A Full MVNO with eUICC, a clear 3GPP NTN hardware strategy, and developing relationships with satellite operators is positioned for the next five years.

The bottom line

O2 Satellite is the most significant development in UK mobile coverage in a decade. The fact that it launched for Samsung smartphone users should not obscure the structural shift it represents. Satellite connectivity is now a standard MNO service. It will become a standard IoT SIM service. The timeline is 2028-2030 for reliable, production-grade DtC capability in the industrial module ecosystem. Plan for it. Don’t wait for it.

11. Europe’s Satellite IoT Market – and Why the UK is Ahead on the Wrong Thing

There is an irony in the UK’s satellite connectivity story that the press coverage has largely missed. The UK is first in Europe on direct-to-cell consumer satellite – O2 Satellite launched on 26 February 2026, ahead of every other European operator. But on commercial satellite IoT for M2M professionals – the standardised 3GPP NTN path that actually matters for industry – the UK is a year or more behind Germany, France, and Sweden.

This is not a minor footnote. It is the clearest possible illustration of the gap between consumer and industrial satellite connectivity, and it tells you something important about where the UK IoT market needs to accelerate.

Deutsche Telekom – the benchmark

While O2 was running internal trials of Starlink DtC in late 2025, Deutsche Telekom was already serving live commercial customers with satellite NB-IoT. DT launched its first satellite-based IoT service with Skylo – a GEO NTN operator – in March 2024. That’s two years ahead of any equivalent UK commercial offering.

In February 2026, DT went further. It launched what it described as the world’s first multi-orbit IoT roaming service, combining GEO satellite connectivity via Skylo with LEO connectivity via Sateliot and OQ Technology – all on commercially available standard hardware. The validated hardware reference is significant: Nordic Semiconductor’s nRF9151 is the first 3GPP-compliant cellular IoT module to support terrestrial NB-IoT/LTE-M as well as NB-NTN over both GEO and LEO. A device using a DT SIM can now roam seamlessly between a terrestrial tower, a GEO satellite, and an LEO satellite – all on the same connection, all on standard 3GPP hardware, today.

In the second half of 2026, DT will add Iridium’s NTN Direct to this portfolio, giving its IoT customers truly pole-to-pole NB-IoT coverage via L-band LEO satellites. By the end of 2026, Deutsche Telekom’s IoT customers will have access to four satellite connectivity options across two orbital regimes. UK IoT SIM customers have access to none, via any provider, on any standardised NTN path.

The broader European picture

Germany is not alone. Telefónica’s o2 Business division in Germany launched a commercial Satellite IoT tariff in March 2026, combining terrestrial and non-terrestrial mobile networks as required for business customers – also via Skylo. This is the same Telefónica brand that owns Virgin Media O2 in the UK – but the German business moved on IoT NTN independently and earlier.

Tele2 in Sweden launched a commercial 3GPP-standardised satellite IoT service via Skylo, making it the first Swedish operator to offer direct-to-device satellite IoT on standardised infrastructure. Orange France has a partnership with Skylo for NB-IoT satellite and is also progressing its AST SpaceMobile collaboration, with demonstrations of voice, SMS, and data planned in Romania during 2026. The European Space Agency and GSMA Foundry are deepening their partnership to accelerate convergence of terrestrial and satellite networks, fuelling NTN standards innovation for 5G and future 6G systems.

On a broader strategic level, the EU’s IRIS² programme, backed by more than €10 billion in public and private funds, aims to deploy a multi-orbit constellation of 290 satellites by the early 2030s – a sovereign European connectivity infrastructure with direct-to-device capability built in from the design stage. This is the EU treating satellite connectivity as critical national infrastructure, not a consumer bolt-on.

Why the UK is behind on IoT NTN

The UK’s position is not through lack of technical capability – it is a combination of regulatory timing, commercial structure, and the fragmented nature of the UK IoT SIM market.

Ofcom only finalised its framework for direct-to-device services in terrestrial mobile spectrum in December 2025, with the licence variation granted to O2 on 25 February 2026. The regulatory environment for satellite NTN in the UK has been slower to mature than in Germany, where the Bundesnetzagentur engaged with satellite IoT roaming arrangements earlier. UK MNOs channelled their satellite attention toward the commercially visible DtC consumer product rather than the quieter industrial IoT NTN path that DT was building out from 2024.

The IoT SIM market structure compounds this. The UK has a large number of Light MVNOs competing largely on price, with limited investment in Full MVNO infrastructure, eUICC platforms, or satellite operator relationships. Deutsche Telekom IoT operates as a globally scaled Full MVNO with direct satellite roaming agreements – a fundamentally different commercial position from most UK IoT SIM providers.

There is also a market size argument. Germany has a substantially larger industrial manufacturing base than the UK – automotive, chemicals, engineering – with correspondingly larger demand for industrial IoT connectivity at scale. That demand pull created the commercial case for DT to invest earlier and more deeply in satellite IoT NTN than any UK operator has yet done.

What this means for UK IoT deployments right now

The practical implication is that a UK-based IoT deployment requiring satellite NTN connectivity today – a remote sensor network, a cold-chain logistics fleet, a distributed utilities monitoring installation – cannot get a UK-native satellite IoT SIM product. But it can access the European ecosystem. DT IoT, Tele2 IoT, and providers building on Skylo’s wholesale platform operate internationally. A UK company can procure satellite NB-IoT connectivity from a German or Swedish IoT SIM provider today, on standardised 3GPP hardware, on commercial terms.

That situation will not persist. The commercial pressure created by O2 Satellite’s consumer launch, combined with the competitive visibility of DT’s multi-orbit IoT service, will accelerate UK MNO and MVNO activity in satellite IoT NTN. Expect the first UK commercial satellite NTN IoT offering to emerge before the end of 2026 – most likely from a provider already in the Skylo ecosystem. When it does, the hardware is already proven, the standards are already written, and the commercial model already exists. The UK market is catching up from behind, but not from far behind.

12. Real-World Performance: What to Expect and What to Distrust

The 150 Mbps headline number attached to O2 Satellite gets quoted constantly. It has almost nothing to do with what the service delivers today. Here is an honest breakdown of current and projected performance across the technologies covered in this piece.

O2 Satellite (Starlink DtC) – current generation

The service launched as a messaging and basic data platform. It is not a broadband replacement and was never described as one. Real-world throughput on the current Gen 1 DtC satellites is low – sufficient for WhatsApp messages, navigation app updates, and safety check-ins, but not for sustained data transfer. No independent benchmarks have been published at the time of writing. The service is correctly categorised as coverage fill-in, not capacity enhancement.

Latency is in the 30-80 ms range due to the LEO orbit altitude of approximately 550 km. This is workable for most M2M protocols – MQTT connections, HTTPS API calls, and basic TCP sessions will function. Applications requiring sub-20 ms latency will not perform well. Geostationary satellite links, by contrast, carry 500-600 ms round-trip delays that break many standard protocols entirely. LEO’s latency advantage over GEO is real and significant for M2M use cases.

Signal strength and indoor penetration

The path loss from 550 km is substantial. The satellite compensates with very high gain phased-array antennas, but the link budget is tight compared to a terrestrial cell. Outdoors with a clear sky view, the service functions as described. Through walls, the signal attenuates significantly – ISPreview noted that the service operates on B3 (1800 MHz), which has better indoor penetration than higher frequency bands, but the satellite’s effective radiated power means indoor reliability is not guaranteed. For IoT devices in metal enclosures, plant rooms, or underground infrastructure, expect the link to be unreliable or absent even where the service nominally covers the area.

Congestion and shared capacity

Each DtC satellite covers a footprint spanning hundreds of kilometres. Every compatible device in that footprint shares the available capacity. Under light load – a remote hillside with a handful of devices – this is fine. Under load – a major terrestrial outage pushing thousands of devices onto satellite simultaneously – capacity compression will occur. Consumer traffic will be deprioritised in favour of emergency calls. M2M background data will sit at the bottom. This is not speculation; it is how QCI-based prioritisation works in every LTE network, and satellite cells are no different.

The 150 Mbps number in context

The 150 Mbps figure comes from SpaceX’s published target for its Gen 2 DtC satellites, planned for launch in late 2027. This is a per-cell peak figure, not a per-user guarantee, and it is a target rather than a demonstrated capability. AST SpaceMobile’s Block 2 BlueBird satellites cite a similar 120 Mbps peak data rate target. Treat these as architectural ambitions for the next generation of hardware rather than a specification for anything you can deploy today.

3GPP IoT NTN – what to expect for sensors

NB-IoT over satellite delivers very low throughput by design – around 250 kbps maximum, and often much less in real deployments. LTE-M (eMTC) over satellite offers slightly more – up to approximately 1 Mbps. Both are more than adequate for sensor telemetry, asset tracking payloads, and periodic status updates. Latency will vary depending on whether a LEO or GEO satellite is used. LEO deployments (Skylo, Sateliot) deliver tens of milliseconds. GEO deployments can reach 250 ms or more, which requires careful protocol selection – standard TCP with aggressive retransmit timers will perform poorly. MQTT with appropriate keepalive settings and QoS level 0 or 1 is a better fit.

13. Frequently Asked Questions

Which phones work with O2 Satellite in the UK right now?

As of launch in February 2026, four devices are supported:

• Samsung Galaxy S25
• Samsung Galaxy S25+
• Samsung Galaxy S25 Ultra
• Samsung Galaxy S25 Edge

That is the complete list. No exceptions, no workarounds. Devices not supported at launch include the entire iPhone range (14, 15, 16, 17), all Samsung Galaxy S21 through S24 models, all Google Pixel devices, and every other Android handset. A phone being recent, flagship, or 5G-capable does not make it compatible. The limitation is modem firmware and baseband chipset support for DtC operation, not signal strength or software.

O2 has stated that support for additional devices will be added over time. iPhone support is widely anticipated – Apple has the engineering capability and has already deployed satellite functionality for Emergency SOS via its own satellite partnership – but no timeline has been confirmed. Samsung’s early position reflects the commercial arrangement struck between Samsung, SpaceX, and O2 ahead of launch, not any inherent technical advantage of Samsung hardware over competitors.

What can you actually do on O2 Satellite right now?

At launch: SMS messaging, WhatsApp, Facebook Messenger, Google Maps, AccuWeather, BBC Weather, AllTrails, X, Yahoo Mail, Samsung Weather, Google Find Hub, and Google Personal Safety. Standard cellular voice calls are not supported at launch. WhatsApp voice calls work. The service activates automatically when terrestrial signal is unavailable – no app, no manual switching required. Think of it as emergency messaging and basic navigation, not general mobile broadband.

Does O2 Satellite work indoors?

Unreliably. The service operates on B3 (1800 MHz), which penetrates building materials better than higher frequency bands – but the satellite link budget is far tighter than a terrestrial cell. A clear view of the sky gives the best results. Through standard glazing the signal degrades. Through thick stone walls, metal cladding, or reinforced concrete, the link will typically fail. For IoT applications in any kind of enclosed structure, do not plan satellite DtC as your primary or fallback link without physical testing.

Will Vodafone get satellite connectivity too?

Yes, via a different technology. Vodafone’s joint venture with AST SpaceMobile – SatCo, headquartered in Luxembourg – is targeting commercial launch in 2026. AST SpaceMobile’s BlueBird satellites use massive phased-array antennas and are positioned as a higher-throughput alternative to Starlink DtC. The two services will compete on different technology platforms: Starlink uses many smaller satellites; AST SpaceMobile uses fewer, much larger ones. Neither service has announced wholesale MVNO access arrangements as of March 2026. For UK IoT deployments, treat the Vodafone/AST service as additive competition that will apply further pricing pressure, not an immediately accessible alternative.

Can IoT routers or M2M modules use O2 Satellite today?

No. O2 Satellite at launch supports only the Samsung Galaxy S25 range. No industrial router, no M2M module, no embedded LTE modem, and no SIM-in-a-box device can access the satellite service today. The modem firmware and chipset requirements for DtC operation have not been incorporated into any IoT module hardware currently available. Module manufacturers including Quectel, Telit, u-blox, and Sierra Wireless have not announced DtC-compatible products. Realistic availability of DtC-capable IoT modules in the market is 2027-2028 at the earliest. Plan accordingly.

Can I use a consumer O2 SIM in an IoT device to get satellite coverage eventually?

Even if satellite module support arrives, a consumer O2 SIM is the wrong tool for industrial IoT. Fair usage policies, CGNAT, the absence of static IP addressing, QoS deprioritisation in congestion, and plastic plug-in form factors all make consumer SIMs unsuitable for professional M2M deployment. This is covered in full in section 7 of this piece. The short answer: it will work briefly, then fail in production for reasons that are difficult to diagnose remotely and not covered by any commercial SLA.

Is satellite IoT just for remote locations?

Primarily yes, for now. The value proposition of DtC satellite is coverage in areas where terrestrial networks do not reach. In areas with good 4G or 5G coverage, satellite adds nothing operationally and introduces the congestion and QoS complexities described above. Satellite as a failover – active only when terrestrial connectivity fails – is the correct architectural model for most M2M deployments. A router or gateway with dual-WAN capability and automatic failover to satellite when terrestrial is unavailable is the sensible target architecture, but the hardware to support it does not exist in the IoT module market today.

Does O2 Satellite cover the Scottish Highlands?

Not at launch. O2 Satellite currently operates only up to the 58th parallel north – roughly a line through Dundee and Stirling. The Highlands, Orkney, Shetland, and the western islands sit above this latitude and are currently outside the service area. O2 and Starlink have said this will improve as next-generation satellites are deployed with different orbital parameters. For IoT deployments in these areas that require satellite connectivity now, 3GPP IoT NTN via providers like Skylo or Sateliot is the more practical near-term option.