Wrecking is not really the right term.
It is causing work for astronomers, and wrecking very few older systems, but generally it is an issue you can work around. I.e. something temporary. What you usually see in my experience of the field is you have some of your work degraded by satellite streaks, which are about 2x more common since starlink, and you understandably complain at starlink. And then get around to coding up a solution to deal with the streaks, spend another few runs until it about works, and eventually forget this was ever a thing.

In more detail, the base issue is, that you are taking an image, with probably minutes or hours or days of exposure, and every satellite passing through that image is going to create a streak that does not represent a star. Naturally that is not good in most cases.
The classic approach here, because this issue has existed since before starlink satellites, is to - depending on frequency and exposure length and your methodology - either retake the entire shot, or throw out at least the frames with the satellite on it, manually.

The updated approach is to use info about satellite positions to automatically block out the very small angle of the sky around them that their light can be scattered to by the atmosphere, and remove this before summing that frame into your final exposure. Depending on methodology, it might also be feasible to automatically throw away frames with any satellite on them, or you can count up which parts of the image were blocked for how long in total and append a tiny bit of exposure only to them at the end.

To complicate this, I think more modern complaints are not about the permanent constellation satellites but those freshly deployed, that are still raising their orbits. Simply because their positions are not as easy to determine, since their orbits are changing. So you need to further adapt your system to specifically detect these chains of satellites and also block them out of your exposures.

The issue here is that you need to create this system that deals with satellite data. And then you need that control over the frames in your exposure, which naturally does not match how exposure used to work in the olden days of film, but to my knowledge does work on all “modern” telescopes.
My knowledge here is limited but I think this covers about 30-40 years of optical telescopes, which should largely be all optical ground based telescopes relevant today. Further, you probably do need to replace electronics in older telescopes, since they were not built to allow this selective blocking, only to interrupt the exposure.

In summary, not affected are narrow fov modern optical telescopes, and in general telescopes operating far from visual frequencies.
Affected with some extra work, would be some older narrow (but not very narrow) fov telescopes, as you now have to make them dodge satellites or turn off shortly, when previously you could have just thrown away the entire exposure in the rarer cases you caught a satellite. This would be software only (not that software is free).
Modern wide fov telescopes might need hardware upgrades or just software upgrades to recover frames with streaks on them.
Old wide fov telescopes may be taken out of commission or at least have their effective observation times cut shorter by needing to pass out on more and more exposure time over satellites in the frame.

It is a problem, yes, but in my understanding one that can be overcome, and is causing the main annoyance and majority of its issues while the number of satellites is increasing, not after they have been increased.
I don’t know of a single area of ground based astronomy that couldn’t be done with even a million satellites in leo.