Replies

Remember Iridium? It was conceived in 1987 by Motorola engineers who wanted a handheld phone that worked everywhere on Earth -- like Starlink is going to do soon. In the 80s, cellular coverage was still sparse, so a constellation of low-Earth-orbit satellites looked like a way to skip towers altogether. The system was announced in 1990, built through the decade, and opened commercially in November 1998: 66 operational satellites, global coverage, and a working mesh in the sky.

The economics never caught the idea. The network cost on the order of $5–7 billion to design, build, and launch on expendable rockets. Handsets ran about $3,000; calls cost several dollars a minute; the phones were bulky and needed a clear view of the sky. Break-even required hundreds of thousands of subscribers. By early 1999 Iridium had only about 10,000. Terrestrial cellular had filled in the cities and travel routes its target customers actually used. In August 1999 the company filed one of the era’s largest bankruptcies.

That was not the end of the hardware. After Motorola nearly de-orbited the fleet, a buyer group acquired the assets for about $25 million, shed most of the debt, slashed operating costs, and rebuilt around government, maritime, aviation, and industrial users. Iridium survived as a profitable L-band niche: small devices, polar coverage, weather-robust links. The original consumer-phone business did not.

Iridium was, in that sense, early. It proved that a LEO constellation could serve the whole planet. What it could not do was pay for that constellation on single-use rockets and a luxury voice tariff. Launch was a one-shot commercial purchase. Satellites were custom and expensive. The design was locked in for years while the ground market moved.

Elon Musk singularly drove SpaceX to change the cost side and to vertically integrate. He treated the booster as reusable hardware, not disposable fireworks, and wrapped rockets, satellites, user terminals, and the network into one company. Falcon 9 landings and rapid reuse made it possible to fly dozens of mass-produced satellites per mission at internal cost. Starlink then used that line to sell broadband rather than voice minutes: home and enterprise internet, ships, aircraft, remote sites. Thousands of satellites, monthly pricing, and iterated dishes replaced a 66-satellite phone network billed by the minute. Starship is the same bet at larger scale.

After Iridium flopped with Moto backing, I figured the whole idea of an economical LEO constellation was dead. Musk's ability to create a vision and execute is astonishing. Whoever would have anticipated one man doing that?

Iridium launched 66 satellites altogether. Starlink today has about 11,000 operational. The U.S. has authorized about 19,408 — a modified first-generation grant of 4,408 plus 15,000 Gen2 satellites. Still waiting on the FCC: SpaceX’s Gen2 application was for 29,988 satellites.

SpaceX filed for a Gen3 / V3 constellation of up to 100,000 satellites in early July 2026. On 18 September 2026 the FCC accepted the application for filing — meaning the paperwork is complete enough to go on public notice and enter review (that is not approval). Gen2 was treated the same way: SpaceX asked for nearly 30,000, and the FCC has so far authorized 15,000, with the rest deferred.