They tried is as a test balloon and, with the new data they got, saw that they can't get the numbers to where they would need to be to be profitable.

The alternative here isn't that they would run something of that size indefinitely, knowing that it will lose them money. It's Google behaving like every single other company currently in existence: deciding up front that it's too risky and not even trying.

At least it created some nice data showing how the legacy ISPs react to competition. Just in case anyone ever needs to prove that this market isn't working.