Which just defers the problem. What's underground will eventually be returned to the surface by [insert geological event here]. What will a landfill be in 10,000 years time?
Deferring the problem seems like a viable strategy in this case. Plastic recycling seems like the kind of thing that technology might have something to say about eventually - especially if the plastic is already conveniently sorted into its own dedicated place. If petroleum demand falls to the point where oil wells shut down, and there's huge piles of plastic ready for the taking just under the surface, the economics of digging it up become more and more favorable. Plausibly, the problem could solve itself in time.