It was a good look then.
Google is making another sexy move with satellites, or so the rumor goes. It’s supposedly via an investment in WorldVu Satellites, which has picked up the spectrum originally assigned to SkyBridge, one of the failed low earth orbit (LEO) ventures of the 1990s. Iridium and GlobalStar actually launched, primarily as voice telephone services, but both had to go through the wringer of bankruptcy before there was even a hope of sustainability.
Technology is better now, but the laws of physics haven’t changed. The selling point of LEO birds is that the round trip is only hundreds of kilometers, instead of more than 60,000 km for geosynchronous satellites, which brings 2 huge benefits: it takes less power to connect and roundtrip latency can be consistent with terrestrial routing times, rather than the half second or so it takes to go up to the Clarke Belt and back, twice.
The downside, though, is that you have to launch several dozen satellites into LEO to get the kind of global coverage that as few as 3 at GEO can provide, or just even the regional coverage of a single one. To keep the economics within plausible ranges (e.g. the $1 to $3 billion Google might be ready to spend), and in keeping with the trade-offs involved in sharing limited spectrum, satellites in a LEO constellation have to be smaller, which means less power, which in turn means less digital bandwidth for each.
There are workarounds, but even the most brilliant technology can only mitigate, not cure, physical limitations. The end result is that low orbital bandwidth is scarce and expensive compared to terrestrial or atmospheric means. It’s not a way to bring YouTube to the masses. Aside from speciality or high value applications, what it is good for is to control balloons and drones and plug the backhaul gaps between them and high capacity ground stations. Creating a heterogenous flock of air and space borne birds would be original, audacious and expensive: much better than resurrecting the grand schemes of 20 years ago.