In this scenario then fees per block could total 1-2 BTC, rather than 0.1-0.2 BTC (example numbers), meaning a price of $80k rather than $800k in your example.
That's just a little less than the last ATH, so it's definitely a feasible amount per Bitcoin. There would also still be a percentage coming from block rewards, at least for the next 30-60 years. And the significance of that could get much higher with increasing value.
I've always had the perspective that the rewards to mining requirements ratio, is a system that can regulate itself, especially as there are so many years between now and the finite supply and many gradual decrements between now and then.