How did connecting phones directly to satellites go from an expensive niche to a feature of mainstream smartphones?
The short answer. in three waves. The first, in the 1990s, built dedicated satellite-phone networks (Iridium, Globalstar, Inmarsat, Thuraya) that worked but failed commercially because terrestrial mobile networks spread faster and cheaper than expected; Iridium went bankrupt in 1999 nine months after launching service. The second, from about 2010, used the surviving networks for niche users (maritime, military, remote industry) and experimented with hybrid satellite–terrestrial networks in the United States (LightSquared, later Ligado, and TerreStar), which mostly failed over interference and financing. The third began in 2022, when Apple's Emergency SOS via satellite and T-Mobile's agreement with SpaceX showed that ordinary smartphones could talk to satellites. What changed was not the physics but three things: cheap launch and mass-produced satellites, very large antennas in orbit, and phone chipsets that could handle the satellite link.
The three waves
| Wave | Period | Systems | Phone | Outcome |
|---|---|---|---|---|
| Dedicated satellite phones | 1990s–2000s | Iridium (1998), Globalstar (1999), Thuraya (2001), Inmarsat handhelds | Special handsets with external antennas | Iridium and Globalstar bankrupt, then restructured; niche survival |
| Hybrid networks and niche markets | ~2008–2020 | TerreStar, SkyTerra/LightSquared (Ligado), DBSD; Iridium NEXT (2017–2019); Globalstar second generation | Special handsets; a few dual-mode phones | Hybrid ventures failed; spectrum later resold; niche markets grew steadily |
| Direct to ordinary phones | 2022 onward | Apple–Globalstar, Starlink direct-to-cell, AST SpaceMobile, Skylo, Tiantong (Huawei), Lynk | Unmodified or lightly modified smartphones | Rapid adoption for messaging and emergencies; capacity race under way |
Sources. company histories and press as reported.
Exhibit 3.1
Each wave failed or changed hands; the spectrum survived and grew in value
- 1998
Iridium service
Bankrupt nine months later
- 1999
Globalstar service
Bankrupt 2002
- 2012
LightSquared collapses
GPS interference
- 2012
DISH buys TerreStar, DBSD
S-band → AWS-4
- 2022
Apple Emergency SOS
iPhone 14 on Globalstar
- 2022
T-Mobile + SpaceX
Coverage Above and Beyond
- 2024
First DTC satellites
Starlink, January
- 2025
EchoStar sells spectrum
~$40B to AT&T and SpaceX
- 2026
Amazon–Globalstar; FCC 15,000
Consolidation and scale
Why the first wave failed
| Factor | Effect |
|---|---|
| Terrestrial mobile spread rapidly in the 1990s | Most potential customers gained cheap tower coverage |
| Handsets large and expensive (~$3,000 for early Iridium phones) | Narrow market |
| Call prices several dollars a minute | Low usage |
| Satellites cost billions (Iridium ~$5B) | Debt could not be serviced |
| No indoor coverage | Poor experience in cities |
Iridium's restructured company bought the assets for about $25 million in 2000, a fraction of their cost, and built a profitable niche business that today earns about $870 million of revenue a year (Chapter 18). The lesson for today's investors is that the first owners of a satellite network may not be the ones who profit from it.
The hybrid-network detour
In the 2000s, US regulators allowed MSS licensees to build an "ancillary terrestrial component" (ATC), so that satellite spectrum could also be used on towers. The idea was a hybrid network: towers in cities, satellites elsewhere. In practice, the satellite part was little used and the value lay in the terrestrial rights. LightSquared's plan for a national LTE network in L-band collapsed in 2012 over interference with GPS receivers; TerreStar and DBSD went bankrupt and their S-band spectrum was bought by DISH (later EchoStar) in 2012, becoming the AWS-4 band that SpaceX bought in 2025 (Chapter 12). The same spectrum has therefore passed through three strategies in two decades.
| Spectrum | Original licensee | Fate |
|---|---|---|
| 2 GHz S-band (US, later AWS-4) | TerreStar, DBSD (ICO) | Bought by DISH (2012); sold to SpaceX (2025) |
| L-band (US) | SkyTerra / LightSquared → Ligado | GPS interference dispute; bankruptcy (2025); AST agreement for L-band use |
| Big LEO L/S-band | Globalstar | Apple partnership (2022); Amazon acquisition (2026) |
Sources. FCC records and press as reported.
What changed by 2022
| Change | Effect |
|---|---|
| Launch cost per kilogram fell roughly tenfold with reusable Falcon 9 (Study S5) | Large constellations affordable |
| Mass production of satellites (Starlink) | Satellites cost hundreds of thousands, not hundreds of millions, of dollars |
| Large deployable phased arrays (AST) | Enough antenna gain to reach unmodified phones (Chapter 4) |
| Smartphone modems capable of satellite modes; 3GPP NTN standard | Phones could support satellite links without bulky antennas (Chapter 22) |
| Mobile operators' interest in coverage marketing | Partners to sell the service |
Lessons from history
| Lesson | Relevance in 2026 |
|---|---|
| Satellite systems are capital-intensive and can fail financially even when they work technically | AST and others depend on continued access to capital |
| Spectrum outlives business models | EchoStar, Ligado, Globalstar spectrum has been resold repeatedly |
| Terrestrial networks are a moving target | 5G coverage expansion shrinks the gap satellites fill |
| Niche markets can be profitable | Iridium's model shows durable value in reliability-focused users |
The second life of spectrum
The most striking thread in this history is the repeated resale of spectrum. Bands licensed for satellite phones in the 1990s and 2000s have passed through bankruptcies, hybrid-network plans and terrestrial ambitions, and in the 2020s have been bought for D2D at prices far above what their original owners paid. The AWS-4 band, for example, began as satellite S-band licences held by TerreStar and DBSD, was bought by DISH in 2012 for about $2.9 billion including the companies' assets (as reported), and was sold to SpaceX in 2025 as part of a roughly $17 billion transaction. Each owner's strategy failed or changed, but the spectrum retained and eventually multiplied its value. For investors, the lesson is that in satellite communications the durable asset is often the licence rather than the network.
What would change this view
If next-generation systems deliver usable voice and data to unmodified phones at prices close to terrestrial plans, the third wave will differ fundamentally from the first two. If usage remains occasional and revenue modest, the history of over-investment followed by restructuring could repeat for weaker players.