Since you bring up crypto as a source of scarcity in storing transactional value, I guess my crankiness compels me to note that there isn't a final say on computational resource asymmetry yet either, and is just as valid of a hope as the Star Trek replicator eliminating physical scarcity or more mundanely as the lessening grip of a diamond cartel.
Right now there's currently a (publicly) huge 'security' margin on the time complexity of hash preimage attacks that allows the mining lottery of generating zero-prefixed hash collisions to proceed orderly for Bitcoin as an example. And while you might be tempted to say that because of the relative stability of BTC as a store of value for the past ~8 years is indicative of just how bulletproof that security really is, you might also realize that a market capitalization of only several billion USD is a drop in the ocean of _positive_ human monetary incentives to exploit, nevermind the alignment of negative incentives that work against anyone actually trying to make use of cryptographic breakthroughs in today's society whether as direct profit within the system as it exists today, revolution, or some other grand long-game.
In the P versus NP debate, our universe _could_ potentially be 'Algorithmica' although I know the majority of research mindshare is heavily assumptive of that being a trivially small possibility.
But there's also research lines on other forms of computational resource asymmmetry that we still assume exists but haven't really come up with proof for yet. For example, besides just deterministic versus nondeterministic polytime and the unproven assumption of one-way functions existing, there's also the open question of NC versus P, which roughly translates to the supposed asymmetry of parallel speedups versus sequential processing -- yet another potential source of scarcity if the current incarnation of crypto puzzles doesn't pan out. Computational space is another potentially unproven source of scarcity: PSPACE versus P.
Even really mundanely there potentially exists inversion 'crypto' puzzles of linear versus quadratic time, while not giving the exponentially wide safety margin every everyone expects for practically, could afford enough clearance to transfer and establish value at least between some highly centralized/colluding resource base against some very poor fraudster if that's the only threat model.
Humanities fascination with ascribing, denominating, and transferring 'value' between each other will be the most enduring and profound research avenues while our species last. And while to me that seems vapid to think that figuring out 'what exactly is money?' is the greatest question in the universe, maybe it really is.