Pretty sure Bill Gates did something similar at the end of the dot-com boom, Google have tried other wide area internet solutions like Loon, Facebook has a similar internet.org project and now Musk.
Previously attempts:
1. We need global internet.
2. Satellites!!!!
3. We need inexpensive rockets to launch our satellites.
4. Fail.
SpaceX / Starlink:
1. We have have excess satellite launching capabilities due to our inexpensive rockets.
2. Global internet!!!
3. Build and launch thousands of satellites.
4. Profit?
Most other satellite internet attempts/successes are in geostationary orbits, so the latency due to the speed of light makes it a non-starter for many situations. Starlink is going to be MUCH closer to the planet in extremely low orbits so latency is well within acceptable ranges for most use cases, and in some situations like cross-continent communication it can actually be faster than ground-based systems (at least theoretically).
The other aspect is that SpaceX is a rocket company who's whole company is based around making launches cheaper. They are probably the only company in history that might be able to pull off launching the multiple thousands of satellites that they require to have a fully functioning system at this altitude.