Non-terrestrial networks (NTN): what they are and how they work

Most of the planet has no cellular signal. Terrestrial mobile networks follow people and power, so they cover cities and highways and stop at the coastline, the desert edge, and the last village on the road. Oceans, poles, mountains, farmland, and shipping lanes sit outside that footprint, and so do the ships, sensors, herds, and vehicles that operate there. Non-terrestrial networks (NTN) close that gap by carrying the same cellular signals through satellites and high-altitude platforms instead of ground towers.

Devices have connected through satellites for years, but only with dedicated hardware on separate networks. In its 2022 Release 17, the 3rd Generation Partnership Project (3GPP), the body that writes the rules for 4G and 5G, folded satellite and airborne links directly into the cellular standard for the first time. A non-terrestrial network is no longer a separate satellite system bolted onto a phone. It is part of the same 5G framework, so ordinary cellular chips and low-power Internet of Things (IoT) devices can reach orbit with the right radio front end.

Non-terrestrial networks (NTN) are wireless networks that use spaceborne platforms (satellites) or airborne platforms (High-Altitude Platform Stations, HAPS) to carry cellular signals to and from areas that terrestrial networks (TN) cannot reach. 3GPP standardized them from Release 17 onward, so an NTN link uses the same 5G New Radio (NR) and cellular IoT protocols as a ground network, not a proprietary satellite protocol.

The platform moved off the ground, but the radio interface stayed the same. A user device talks to a satellite the way it would talk to a cell tower, and the network treats the satellite as another node in the radio access network. A device built for NTN holds its connection as it moves from a terrestrial cell into open ocean or remote land where no tower exists, using one modem and one standard rather than two separate systems.

User equipment (UE), meaning the phone, module, or sensor, transmits a cellular signal upward. A satellite or HAPS platform receives it and relays it to a ground station, called a gateway or feeder link. The gateway connects to the core network on the ground, which routes the traffic to its destination the same way any mobile call or data session is routed. The return path runs in reverse.

The distances are enormous, which stretches signal timing and weakens the received power, and in low orbits the satellite moves fast relative to the device, which shifts the signal frequency. NTN standards solve both with satellite ephemeris data and Global Navigation Satellite System (GNSS) positioning in the device, so the UE can pre-compensate for delay and frequency shift before it transmits. How the satellite handles the signal once it arrives depends on the payload.

The payload is the communications equipment a satellite carries. How much work it does on the signal, beyond receiving and resending it, is what separates the two designs.

A transparent payload, also called bent-pipe, does the minimum. It receives the signal, converts its frequency, filters it, amplifies it, and sends it back down. It does not decode anything. The base station stays on the ground, and the satellite acts as a mirror in the sky.

Transparent payloads are cheaper, simpler, and compatible with existing terrestrial network equipment, so 3GPP Release 17 specified them first. The trade-offs are higher end-to-end latency and a dependence on large, well-placed ground stations.

A regenerative payload puts processing on the satellite. It demodulates and decodes the signal, and can route it on board, effectively placing a base station in orbit. This lowers latency, reduces the load on ground infrastructure, and enables inter-satellite links, where satellites pass traffic to each other before reaching a gateway. Regenerative payloads are the direction of travel for low-orbit 5G NTN, and 3GPP Release 19, frozen in December 2025, moved the standard decisively in that direction.

Altitude sets the character of a non-terrestrial network. A higher orbit sees more of the Earth from one satellite but adds distance, and distance adds delay and signal loss.

• Geostationary orbit (GEO) sits at roughly 35,786 km. A GEO satellite appears fixed over one spot, so three of them cover almost the entire populated globe, and antennas can point at a stationary target. That distance costs around 240 to 280 ms of one-way latency, too slow for tightly interactive traffic but fine for messaging, broadcast, and delay-tolerant IoT.
• Medium Earth orbit (MEO) runs roughly 8,000 to 20,000 km up, with one-way latency near 40 to 70 ms. It is a middle ground, covering wide areas with a modest number of satellites at moderate delay.
• Low Earth orbit (LEO) sits at roughly 500 to 2,000 km, with one-way latency of a few milliseconds and round-trip times in the tens of milliseconds. LEO gives the responsiveness that broadband and real-time services need, but each satellite covers a small area and moves quickly overhead, so LEO needs large constellations of hundreds or thousands of satellites and has to manage strong Doppler shift and frequent handovers.

HAPS platforms, such as stratospheric balloons and solar aircraft, fly at around 20 km and act as very low, very local NTN nodes for targeted coverage. For the 6G era, very low Earth orbit (VLEO) and highly elliptical orbit (HEO) are under study to push latency lower or hold coverage over high latitudes.

A non-terrestrial network comes in two technology families, built for two different jobs. 3GPP Release 17 introduced both in 2022, and Release 18 (the start of 5G-Advanced) enhanced them with wider bands and better mobility. One family is for small low-power devices sending a little data, the other for richer broadband and voice.

Both operate in dedicated satellite bands. Band n255 sits in L-band, with an uplink of 1626.5 to 1660.5 MHz and a downlink of 1525 to 1559 MHz. Band n256 sits in S-band, with an uplink of 1980 to 2010 MHz and a downlink of 2170 to 2200 MHz. Both use frequency-division duplex. These are standardized mobile-satellite bands, distinct from the higher Ka-band and Ku-band that Release 18 added for broadband terminals on aircraft and ships.

“Direct-to-cell” or direct-to-device (D2D) usually refers to an unmodified consumer smartphone connecting to a satellite for texts or basic data. That is a specific consumer application of NTN, not a separate technology.

NB-IoT NTN, sometimes written IoT-NTN, brings Narrowband IoT (NB-IoT) and LTE-M to satellite links. It targets the same devices as terrestrial cellular IoT, meaning small, battery-powered, low-cost modules that send short messages rather than streams. Think of a sensor in a field, a tag on a shipping container, a meter on a pipeline.

The data rates are low and the traffic is delay-tolerant, which suits higher orbits and simple transparent payloads. NTN IoT is already in commercial service, with operators running Release 17 NB-IoT NTN over existing satellites today.

5G NR NTN brings full 5G New Radio to satellite. It carries higher data rates, supports voice and interactive traffic, and covers both smartphones (direct-to-device) and higher-end terminals. Because it needs lower latency, 5G NTN leans toward LEO constellations and, increasingly, regenerative payloads that put a base station in orbit. Most of the near-future investment sits here.

A non-terrestrial network reaches the oceans, the poles, and remote land that no ground network serves, bringing large populations and a great many off-grid assets into cellular range for the first time.

When a terrestrial network fails in a disaster, an emergency, or a backhaul outage, an NTN link keeps devices connected or provides failover. That continuity matters for public safety and for critical infrastructure that cannot afford a silent period.

Reaching a remote asset with a private radio network or a dedicated satellite terminal is expensive. NTN lets a standard cellular device do the job, which lowers the cost of connecting things spread across wide, empty geography. Because NTN is now part of the cellular standard, a device can reach a satellite directly, without separate satellite-phone hardware, using one modem that roams between ground and space.

Distance drives most of the difficulty. Long propagation delay raises latency, especially in GEO, which rules some orbits out for interactive use. Fast-moving LEO satellites impose a large Doppler shift on the signal frequency, which the device must correct. Path loss over hundreds or thousands of kilometers leaves a thin link budget, so every decibel of antenna and front-end performance counts.

Handover has to work not only between terrestrial and non-terrestrial networks but between satellites that are themselves moving across the sky, and the network must manage timing advance to keep uplink transmissions aligned despite the changing distance. Spectrum and regulatory coordination across national borders is a further constraint, since a satellite footprint does not stop at a country line.

The device itself is easy to underrate. The terminal runs on a tight power budget, and it needs an efficient, omnidirectional NTN antenna that works across the L-band and S-band satellite bands while a satellite could be anywhere overhead. Fitting that antenna performance into a small, low-power enclosure is a real design problem, one item among several rather than an afterthought.

Non-terrestrial networks earn their place wherever devices operate beyond terrestrial coverage:

• IoT and asset tracking: Off-grid and cross-border tracking of containers, trailers, and equipment that spend part of their journey outside cellular range. This is the clearest early NTN IoT market.
• Maritime and logistics: Vessels, offshore assets, and cargo moving across oceans where no tower reaches.
• Agriculture and livestock: Soil, water, and animal sensors spread across farmland far from infrastructure.
• Energy and utilities: Pipelines, grid equipment, and remote sites monitored across wide, empty terrain.
• Aviation and automotive: In-flight connectivity and connected vehicles that need coverage on remote routes, including direct-to-device links.
• Emergency and public safety: Connectivity that survives when the ground network is down, and messaging from areas with no coverage at all.

Broadband cases grow as 5G NR NTN and regenerative payloads mature, adding voice and higher-rate data to the list. The terrestrial side of the same low-power picture comes down to the difference between LTE-M vs NB-IoT and the fundamentals of cellular IoT connectivity.

For a connected device, NTN IoT means a cellular IoT product can stay online when it leaves terrestrial coverage, using the same modem and much of the same design as its ground-only version. The value shows up in exactly the deployments that already stretch cellular IoT hardest, like global IoT asset tracking across regions and borders, where a tracker has to work on many bands and hold a link wherever it travels.

Designing for it puts pressure on the power budget, because the device is often battery-fed and far from service; on band coverage, because it now has to reach the L-band and S-band satellite bands on top of its terrestrial bands; and on the antenna, because it needs efficient, omnidirectional radiation across that range in a small enclosure, with a satellite that could sit anywhere overhead.

A non-resonant antenna is built for exactly that combination. Rather than a resonant element cut for a fixed band, the Virtual Antenna® approach uses a small antenna booster component matched to the required bands through an external matching network, so a single component covers a wide frequency envelope, from roughly 400 MHz to 10 GHz. That envelope already spans the cellular and GNSS bands an asset tracker relies on today. Because the matching network is what selects the bands, retuning toward the NTN L-band and S-band window is a matching change, not a new antenna or a PCB respin.

Ignion has no NTN-band test data yet, so this is band coverage by architecture, not a measured NTN result. What is proven is terrestrial, where the same component ships in multi-band LTE-M and GNSS trackers across 200+ country operators. Ignion’s hardware designer’s guide to cellular IoT antenna design covers how the matching approach works in practice.

Proof from real cellular IoT and asset-tracking deployments:

• How Velavu built a global tracker on a single antenna
• Every asset-tracking protocol from one antenna

3GPP Release 18 widened 5G NTN to higher bands and improved mobility, and Release 19, frozen in December 2025, tipped the payload debate toward regenerative designs and added store-and-forward operation for delay-tolerant IoT that does not need a live feeder link. Very low Earth orbit and highly elliptical orbit are under study to cut latency further and hold coverage at high latitudes, and direct-to-device is broadening from niche messaging toward mainstream support.

The larger shift is architectural. In 5G, NTN was added to an existing standard. In 6G, satellite and airborne access are expected to be native from the first specification, with terrestrial and non-terrestrial networks treated as one continuous fabric. Release 20 is set to bundle 5G-Advanced NTN features with early 6G study, and Release 21 is expected to carry the first normative 6G work. A device sold in that era should reach a network whether or not a tower is in range, and the design question becomes how to build hardware ready for both.

Non-terrestrial networks are wireless networks that use satellites or high-altitude platforms to carry cellular signals where ground networks cannot reach, such as oceans and remote land. 3GPP standardized NTN from Release 17, so they use the same 5G and cellular IoT protocols as terrestrial networks.

A device transmits a cellular signal to a satellite or HAPS platform, which relays it to a ground gateway that connects to the core network. The device uses GNSS positioning and satellite ephemeris data to pre-correct for the long delay and frequency shift that distance and satellite motion introduce.

A transparent (bent-pipe) payload only converts, filters, and amplifies the signal, keeping the base station on the ground. A regenerative payload decodes and routes the signal on board, acting as a base station in orbit. Regenerative payloads cut latency and enable inter-satellite links.

NTN use geostationary orbit (GEO, ~35,786 km, highest latency), medium Earth orbit (MEO, ~8,000–20,000 km), and low Earth orbit (LEO, ~500–2,000 km, latency in milliseconds). Higher orbits cover more ground per satellite; LEO gives lower latency but needs large constellations.

NB-IoT NTN carries low-power, delay-tolerant IoT traffic from small battery devices, and is already in commercial service. 5G NR NTN carries broadband data, voice, and direct-to-device links for smartphones and richer terminals. Both entered the standard in 3GPP Release 17.

The main advantage is coverage everywhere, including oceans and remote land, plus resilience when ground networks fail. The main challenges are latency from long distances, Doppler shift from fast satellites, a thin link budget, handover across moving satellites, cross-border spectrum coordination, and the device power and antenna budget.

Direct-to-cell, or direct-to-device, is a specific use of NTN where an unmodified consumer smartphone connects to a satellite for messaging or basic data. NTN is the broader standard covering that case plus IoT modules and higher-end terminals across several orbits and bands.