On human timescales, it's basically forever. Hopefully we'll develop the tech to clean up debris in space, but it's extra challenging to do it in geostationary orbit since it's so far away from Earth, both in terms of actual distance, and delta-V.
> People worry about LEO constellations causing Kessler syndrome, but the reality is that LEO debris deorbits in the order of months/years.
It's a little more complicated than that. The time to spontaneously deorbit is based on orbital height. Starlink can deorbit on its own in 5-10 years because it's orbiting so low. But any OneWeb satellites that malfunction[1] will take 1000+ years to deorbit because they're up at 1000+ km.
---
1. Like this one
https://spacenews.com/oneweb-mulls-debris-removal-service-fo...
Rendezvousing is pretty established tech so long as you know a precise and stable orbit of your target, afaik? Which, for geo, we would I think. So taking up some grabbing mechanism probably does it, then use ion engines, burning retrograde (avoiding the need for heavy fuel) until you get it to a low LEO orbit, let loose, and let the problem solve itself within a few ~weeks. Then move on to the next piece, so you don't need a launch to orbit for every individual piece of debris. You also don't have to circularise your orbit to just rendezvous and grab it. And you probably also don't have to go out of plane even if the target object is, if I'm visualising this correctly, because there's always a node where the planes intersect and you can just start the path up to geo at the right point in your LEO orbit
Grossly simplified, devil in the details, but this seems very possible with today's technology and potentially less expensive in terms of delta V than it may seem at first glance