Bitcoin cannot depend on altruism to survive, its success is contingent on network incentives aligning with what is needed to keep the network operational. So I think it is not really worth considering that as an option. You could perhaps argue that NACs are incentive-driven, in that the user is incentivized to see the network that they store their value in maintain an acceptable level of security, but the prudent individual is going to find a financial instrument that gives them the best guarantees for the cheapest price. From my understanding of them, it isn't clear that depending on altruistic NAC payments constitutes a low-cost and incentive-aligned method of maintaining security.
I also doubt that any soft fork that enforces asymmetric fees based on UTXO age/value would ever be widely accepted/adopted (consider: soft fork adoption at least somewhat depends on an 'economic majority' adopting the change, so would you expect large holders to adopt a fee-policy that disproportionately affects them?).
Anyways, it isn't clear that there is a problem in the first place. The Bitcoin network is self-balancing, there is no 'minimum required hashpower' to keep it functioning. Less hashpower perhaps inspires less confidence in the network's security, and thus the network will be seen as less valuable, but it will continue to function nonetheless. Perhaps the chain tip will be less stable than it is now, but as long as the incentives continue to work, the user can have an increasing amount of confidence in the finality of a confirmed payment as each new block comes to pass.