Non-Terrestrial Networks: How Satellite IoT Stopped Being a Niche and Became a Standard
For years, connecting an asset in a cellular dead zone meant a bulky, proprietary satellite terminal and a separate bill. 3GPP has quietly changed that. Here is what NTN actually is, where the standard stands in 2026, and why industrial IoT buyers should now assume every remote device can talk to space.
The problem NTN was built to solve
Cellular coverage maps flatter the truth. They show the places people live, not the places industrial assets work. Pipelines, remote substations, agricultural sensors, long-haul fleets, offshore buoys and rail-side equipment routinely sit beyond the edge of any terrestrial tower. By most estimates, terrestrial mobile networks reach only around 15 percent of the planet’s surface, which leaves the overwhelming majority of the Earth, most of it ocean, desert, mountain and farmland, without a signal.
Until recently, bridging that gap meant bolting on a dedicated satellite terminal: expensive hardware, its own antenna, its own airtime contract and its own integration headache. That economic and engineering overhead is exactly why so many high-value remote assets were left running on manual check-ins, store-and-forward workarounds, or no telemetry at all. Non-Terrestrial Networks remove the overhead by folding satellite connectivity into the same cellular standard the device already speaks.
What “non-terrestrial network” actually means now
NTN is not a single product or a single constellation. It is a set of 3GPP specifications that let a standard cellular device reach a satellite using the same air interface it uses to reach a ground tower. The turning point was 3GPP Release 17, which for the first time added native satellite support to both narrowband IoT and full 5G New Radio. The significance is architectural: instead of a separate satellite radio stack, chipset vendors can extend existing cellular silicon to talk to space, which is what collapses the cost premium that historically made satellite IoT uneconomic.
That single decision reframes satellite from an exotic add-on into a coverage layer. One air interface, one ecosystem, and eventually one modem that does not know or care whether the network it is talking to is on a mast or in low Earth orbit.
The two families: IoT-NTN and NR-NTN
The most common source of confusion is treating NTN as one thing. Release 17 actually produced two related but distinct tracks, and they serve very different devices. IoT-NTN carries low-power sensor telemetry over NB-IoT and LTE-M. NR-NTN carries broadband and, increasingly, direct-to-smartphone voice and messaging. Get these two mixed up and every downstream assumption about throughput, power and cost goes wrong.
| Dimension | IoT-NTN (NB-IoT / LTE-M over satellite) | NR-NTN (5G NR over satellite) |
|---|---|---|
| 3GPP origin | Release 17 (LTE-based) | Release 17 (NR-based) |
| Target traffic | Small, infrequent sensor payloads | Broadband data, voice, direct-to-device |
| Typical throughput | Roughly 20 to 60 kbps, NB-IoT class | Mbps-class, scaling toward broadband |
| Device type | Trackers, meters, remote sensors | Smartphones, gateways, higher-tier terminals |
| Power profile | Ultra-low, multi-year battery life | Higher draw, mains or large battery |
| GNSS requirement | Mandatory for NB-IoT NTN | Assisted, implementation dependent |
| Best fit | Agriculture, utilities and asset tracking in dead zones | Emergency messaging, D2D voice and text, broadband fill |
Direct-to-device: the commercial wave breaking now
If Release 17 was the standards milestone, 2026 is the commercial one. Direct-to-device, sometimes branded direct-to-cell, lets an ordinary handset or module connect straight to a satellite with no special dish. The regulatory groundwork landed when the US FCC approved its Supplemental Coverage from Space framework in late 2024, formally letting mobile operators partner with satellite networks to extend their own licensed spectrum from orbit.
The operator announcements followed fast. T-Mobile went first in the US with its SpaceX-backed service; AT&T and Verizon are moving to launch their own direct-to-device offerings through AST SpaceMobile; and the pattern is repeating internationally with Vodafone, Rakuten, Orange, Virgin Media O2 and Airtel among the operators lining up satellite partners. Analysts tracking the space describe 2026 as a potential tipping point where D2D shifts from early US and China deployments into mainstream global availability, with enterprise use in utilities, oil and gas, maritime and agriculture driving the near-term IoT demand rather than consumers.
Reality check on the numbers. The momentum is real but uneven. Of the roughly 130 telco-satellite collaborations announced since 2024, only about a fifth had reached a beta or commercial launch by late 2025. Satellite is arriving as a genuine coverage layer, but “announced” and “live in your region” remain two very different things.
How hybrid failover actually works
For industrial deployments, the interesting design pattern is not satellite-only. It is the hybrid gateway that treats terrestrial and non-terrestrial as one continuous coverage fabric. The device prefers 4G or 5G when a tower is in reach, because it is cheaper and faster, and falls back to satellite only when terrestrial signal drops. Done well, the switch is automatic and the application layer never needs to know which bearer is carrying its data.
Release 18 hardened exactly this behaviour, adding coverage enhancements and the mobility procedures needed to hand a device cleanly between terrestrial and non-terrestrial networks. Release 19 pushed further still, introducing regenerative payloads that turn a satellite into a base station in space rather than a simple relay, and formalising store-and-forward operation. For intermittently connected IoT, store-and-forward is quietly transformative: a device can hand its data to a passing satellite, which holds it in orbit and drops it to the ground network when a feeder link becomes available. Continuous coverage stops being a prerequisite.
The right architectural assumption for any new remote design is therefore dual-mode by default. Assume the device will sometimes see a terrestrial network and sometimes a satellite, and that it will not know which in advance. Release 17 was written around precisely that assumption.
The hardware reality: what dual-mode costs you
Because NTN reuses the 3GPP air interface, module vendors can add satellite support to existing cellular chipsets, which is what keeps the bill of materials conversation manageable rather than alarming. Integrated parts that combine terrestrial cellular and NTN are already sampling and shipping, and for many designs the incremental cost of choosing an NTN-capable module now is small next to the cost of a hardware redesign later.
There is one constraint worth planning around early. NB-IoT NTN devices must carry a GNSS receiver, because the network relies on the device knowing its own position to pre-compensate for the enormous Doppler shift and delay of a fast-moving LEO satellite. That changes your power budget, your antenna planning and your time-to-first-fix assumptions. It is not a dealbreaker, but it is a line item that pure terrestrial designs never had to carry.
The performance envelope, roughly. Expect NB-IoT NTN downlink in the region of 20 to 60 kbps under good conditions, one-way latency around 20 to 40 ms over LEO, and closer to 270 ms over GEO. Plenty for telemetry, alarms and periodic reporting. Not a pipe for firmware-over-the-air or video.
Where the standard is heading
NTN is maturing release by release, and the trajectory matters for anyone specifying hardware on a three-to-five year horizon.
| 3GPP release | Status | Key NTN additions |
|---|---|---|
| Release 17 | Frozen | First native satellite support for NB-IoT / LTE-M (IoT-NTN) and 5G NR (NR-NTN); Doppler and delay compensation |
| Release 18 | Frozen | Coverage enhancements, terrestrial-to-NTN mobility, higher-frequency and Ka-band study |
| Release 19 | Complete | Regenerative payloads (base station in space), store-and-forward, device-to-device via satellite |
| Release 20 | In study | Voice over satellite, multi-orbit operation, enhanced location and emergency services |
The honest caveats
Before you rip out a terminal. NTN is early. Compatible devices and live satellite coverage are still limited and regionally patchy, throughput is modest by design, GNSS adds power and antenna overhead, and many headline operator partnerships remain at the announcement stage rather than in service. The sensible posture for most fleets is to design in NTN capability now so no redesign is needed later, while treating satellite as a coverage insurance policy rather than a primary bearer until the ecosystem in your specific market has matured.
What this means for industrial IoT: coverage is only half the question
NTN answers one of the two questions that define a modern connectivity decision: can this device get a signal at all, anywhere it might travel? For assets that roam beyond the tower line, the answer is now increasingly yes, and the honest planning assumption is to build dual-mode by default.
But coverage is only half the picture. The other half is capability: given that a device can connect, what tier of cellular does it actually need? That is where the debate is shifting from NB-IoT and LTE-M at the bottom, past LTE Cat-1 bis in the middle, to 5G RedCap and its lower-power variant eRedCap. NTN extends coverage outward; RedCap fills the capability gap in the middle. Together they describe the connectivity stack industrial IoT will be specified against for the rest of the decade. We break down the capability side of that story in our companion guide to 5G RedCap and eRedCap.
Sources and further reading: 3GPP NTN overview; Ericsson Technology Review on satellite direct-to-device; STL Partners Skywatch D2D tracker; GSMA Mobile Policy Handbook on D2D; US FCC Supplemental Coverage from Space framework. Technical figures for throughput and latency are indicative and vary by constellation, orbit and conditions.
