The highest numbers I can find for the final Starlink constellation is 40,000 satellites. Let's assume Starlink and its competitors have constellations totaling 100,000 satellites, and satellites need to be replaced every five years, and each satellite weighs 1 ton. That means 20,000 tons of vaporized satellites per year. The atmospheric emissions would be 3.88 parts per billion per year. This would still be less than the mass of asteroids and space dust that burn up in the earth's atmosphere every year.

If the reentering satellites were somehow transformed entirely into chlorine gas that somehow stayed in the atmosphere forever, we would reach the OSHA permissible exposure limit of 1ppm after 250 years. Chlorine is detectable by smell at 3ppm, which would take 750 years.

It's very likely that the vast majority of the vaporized satellites are inert, as they are basically incinerated on reentry. It's also likely that most of of the vaporized satellite does not stay in the atmosphere for very long. The only way this could be a problem is if the satellites emit a long-lived compound that catalyzes a reaction in the atmosphere, similar to how CFCs destroy the ozone layer. So far, the only candidate for that is aluminum oxide particles, and solid rocket boosters create more of that than reentering satellites. (Fortunately aluminum oxide isn't nearly as bad for the ozone layer as CFCs, and SpaceX does not use solid boosters.)

Also once you are launching tens of thousands of tons to orbit per year, it starts to become feasible to build infrastructure in space. Satellites at the end of their service life contain valuable raw materials. It would likely become cheaper to refurbish or recycle them rather than deorbit and launch new ones.