Non-terrestrial networks: what the satellite connects, and what stays on the ground

Why does a satellite network need ground stations, and what does it mean when one is ready? A Telesat Lightspeed station in France, announced by Orange and Telesat on September 17, 2026, offers a concrete example. Ground infrastructure is advancing while the constellation remains in development. Confusing those stages would turn the commissioning of one station into apparent availability of the entire network, overstating what the announcement actually establishes. [1]

What NTN means

NTN stands for non-terrestrial network. In mobile telecommunications, it describes networks where a space-based or airborne platform participates in radio access. It is neither the name of a constellation nor a universal guarantee of compatibility with any phone. Technical literature on adapting 5G to satellites distinguishes necessary changes in access, architecture, services and terminals: integrating a space link affects the complete system rather than just one antenna. [2]

It is also useful to separate access from transport. With direct access, a terminal establishes its radio link with the satellite. In a transport or backhaul application, a satellite can connect a remote installation to the core network while users access another network locally. Research on satellite 5G for IoT describes this second possibility. Sharing the word “satellite” does not make these architectures the same product or require identical devices. [3]

The French example

At Bercenay-en-Othe, Orange hosts a gateway: a station that will connect the constellation to ground infrastructure. According to the joint release, Orange provides connectivity toward Telesat's planned point of presence in Paris. A point of presence is a location where a network interconnects and hands traffic to other infrastructure. The station and transport connection are parts of an end-to-end path, rather than substitutes for satellites and terminals. [1]

In that announcement, Telesat placed its anticipated global in-service date in the first quarter of 2028. This is the company's expectation, rather than verified availability. The same source presents security, resilience and low latency as design objectives. TelcoBrief's news article about the station preserves that distinction: the commissioned component is ground infrastructure, while the project's global capabilities are not demonstrated by that event. Readers should not infer otherwise from the word “commissioned.” [1]

Orbit changes the problem

A low Earth orbit places a satellite closer to Earth than a geostationary orbit. Distance affects signal propagation time, but total delay also includes processing and travel through the rest of the network. Relative motion also requires handling changes in the link and synchronization. “Low latency” therefore requires a complete measurement and its conditions; it cannot be inferred solely from the name of the orbit being used. [2]

One example of experimental evidence comes from the European Space Agency. In December 2024, it reported a Ka-band 5G NTN connection between its ESTEC laboratory in the Netherlands and Telesat's LEO 3 demonstration satellite. ESA said it maintained the link during the satellite pass using Amarisoft technology and 3GPP standards. That documents a specific test, rather than showing that any commercial mobile phone could use that particular connection. [4]

ESA participates in the experiment and Telesat supplies the satellite: they are primary sources about their own work, rather than an independent audit of a service sold to the public. Nor should the LEO 3 result automatically be transferred to Lightspeed. A demonstration helps reduce a technical uncertainty; by itself, it does not measure commercial capacity, continuous coverage, prices or experience under load for a different constellation. [4]

Coverage does not mean unlimited capacity

A paper by Mikko Majamaa and five coauthors, released as a preprint in May 2024, assessed satellite IMT-2020 requirements through NR-NTN simulations involving a low Earth orbit satellite. It examined throughput, capacity per area and spectral efficiency with particular antenna and frequency-reuse configurations. Its results depend on those assumptions. The methodological lesson is to ask how available resources are shared, in addition to asking whether a signal reaches the location. [5]

For a business seeking to connect remote installations, these distinctions lead to practical questions: what terminal is needed, which traffic can be carried, with what continuity, and which parts of the path are included in the service? These are evaluation criteria, rather than capabilities TelcoBrief attributes to a specific offer. The sources reviewed do not support calculating a universal cost or claiming that one architecture will replace all terrestrial alternatives.

What to look for in the next announcement

The useful check is to identify both component and status: a station installed, a link tested, satellites available or a service contracted. Then look for compatible terminals, coverage conditions and end-to-end measurements. NTN expands connectivity options, but each architecture retains its own constraints. Orange and Telesat's example shows precisely why understanding a space network also requires following what happens on the ground, rather than treating every milestone as proof of a complete commercial offering. [1][2]