The first rocket may take off sooner than 2040. But Starlink is not just a rocket, it is a complete business process, with a launch regularity and price. A Starlink satellite's worth of space on a Falcon 9 costs 500k-750k. With about ten thousand satellites, which last about five years, this means maybe a billion and a half per year spent on the space arm of the business, not counting ground stations. If they had to spend, say, ten times this, Starlink wouldn't be profitable today. And that's pretty much reality: the Ariane rocket costs ~$100m to Falcon's ~$15m (nobody knows what Zhuque-3 costs); I think cost per kg is 5000 vs 900. You could get it down to ~1.5B a year by narrowing it to just the latitudes overhead the EU, but then you cut the potential revenues too and have the same problem.

> but then you cut the potential revenues too and have the same problem.

How many starlink clones are there really customers for?

Many people have fiber, and in an urban area you'll probably prefer 5G, if you can't get fiber or wired internet.

Starlink is great if you live in the middle of nowhere, but few people do.

Even if you could do a competitive launch cost, the number of customers is limited.

Sure but the Chinese military can easily afford that.