Because you are not looking. The fact that spacex is having to spend engineering time at all making improvements to operate in an actively jammed environment means software and possibly hardware engineering resources not being spent on scaling the network, working on their next gen constellation, etc. For a company that runs as lean as SpaceX on engineering headcount (based on reading Blind and Glassdoor reviews), that opportunity cost is easily tens of millions of dollars.
You’re also drastically misunderstanding the lifetime and duty cycles of LEO sats. Those things have to have time to dump heat and charge their batteries. Serving in Ukraine means worse coverage in America 90 mins later. It’s not just a magic mirror in the sky that costs nothing to cover the ground.
The fact is: the Russians have to date been unsuccessful in suppressing Starlink signals, for a variety of reasons including some countermeasures.
But let's not pretend that the main cost is in that aspect of operating the service. The main cost is in making and launching the satellites and once those costs have been sunk further use of the system assuming the terminals are bought and paid for.
As time goes on and parties are getting more inventive likely there will be some more defenses against countermeasures. But a LEO based satellite system has a number of advantages over a GEO based one and those operate in war zones all the time.