Looking at this and Google Loon, I think Google Loon (as far as I've seen) is a flop compared to this, given that this works to some extent. Google Loon's capabilities also seem subpar to this widespread network.
I worked on Loon for a few years. It's a different use case. At "only" 20km altitude, they are able to provide LTE service direct to standard handsets. Starlink requires a dedicated rooftop terminal with a phased array antenna.
From what I remember about Google Loon was that it could be deployed in areas after a disaster. That way any infrastructure that was unaffected by the disaster would not be brought down due to the influx of people trying to communicate to those in and out side the immediate area.
Interestingly, Google is believed to be the largest outside investor to Starlink.
They decided to hedge their bets?
It is typical for Google to fund multiple competing projects, whether internal or external, for almost anything the founders find interesting.