Regardless of whether or not a deflationary end-state is better or worse, it seems non-controversial that the transition from our current situation to a deflationary one would be absolutely brutal. We're talking about turning the global economy on it's head.
It won't help how choosy investors are, or how well the economy performs. The money supply is fixed. So for someone to make 120 bitcoins from an investment of 100 bitcoins, someone else has to lose 20 bitcoins, which would be other investors, or the public (workers / consumers).
We currently have many currencies, if Bitcoin is so absurdly better, then maybe there would be a transition. Otherwise I'm sure many currencies will be used based on the application.
Bitcoin seems to be best for hording and spending. It holds value and can be transferred easy and cheap.
You can still have companies using currencies that incentivise loan/investing. Bitcoin currently exists and the loans keep being written.
I think as bitcoin grows its going to become big and boring. No 10x gains, not even 2x gains, like 10% swings over the course of a year. It wont be fun, it will be a store of value if you dont trust fiat. Right now I think its crazy undervalued which is causing the explosion of crypto. As BTC gets bigger, investment will start to look better as it provides better returns.
It can't get any bigger, the transaction limits, the "halvings", the liquidity issues after over a decade of operation, so I guess that experiment has run its course.
Ergo, there must be a hole in your theory somewhere.
Without inflation, i.e. fixed money supply, it would be impossible for a majority of loans to be repaid with interest, or give a positive return to investments. Note that this fact is independent of how well the borrower or business performs. (This is like a zero-sum game).
So, it would be irrational for investors or lenders to part with their deflationary money, as the net profit is always zero (or negative because of other costs).
I am very interested to know if there is a way around this problem.