For the reward of mining to remain commensurate with the current value (in USD), i.e, 0.15BTC giving miners about the same amount they are earning now, which is ~$125k, then the value of 1 Bitcoin needs to rise to above $800k. That's a huge leap from the current price levels.
This is based on the assumption that total fees per block will remain static, which as I mentioned above, I don't think they necessarily will.
We already know that block space is too limited as it stands to scale bitcoin to global usage, which is why second layer solutions are being developed. Currently, almost all of my transactions pay a fee of <5 sats/vbyte. If I'm paying for a coffee, fast food, or some other small transaction, then a fee much higher than that starts to become a significant amount of the transaction and therefore unacceptable to me. Let's say, however, that I move all my coffee buying on to Lightning. Now I can pay a fee of 50 sats/vbyte to open a channel, and then buy 100 coffees through that channel before I need to close it. In this scenario I have saved money on fees by paying 0.5 sats/vbyte per each of my 100 transactions, but at the same time I have paid 10x what I would usually pay for the block space.
So provided there is enough demand for block space to support things like opening/closing Lightning channels at a higher fee rate (which is still economical for the end user), as well as very large transactions which will pay a higher fee for the added security of being on the base layer, then the fees paid could increase by an order of magnitude or more, while individual users still pay less by moving most of their transactions on to Lightning.
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.