This "reclamation", on a blockchain such as implemented BTC currently, has no privkeys and cannot "reclaim" or "steal" any coin to any address, it can just erases them from the chain without further issuance. Call it erasing instead, a dick move since... The expectation of BTC and most other schemes of blockchains to date from their inception is that you don't need to in any way mark current, refresh, or forward activity spend any transaction, and that old contracts multisig txs etc always remain valid over time. Users with value on those chains, whether bound in contracts or owned, will refuse and revolt against any such invalidation attempts, or migrate value through exchange to other coins when that is the only cheaper option to such an attack. The time to invent such lossy or activity requirements is in new coins that set such expectations of dynamism to their users at launch. Storage... though density and compute may continue "Moore's", preparing for if it doesn't, or simply to optimize and reduce consumption and carriage, seems prudent where available and sensible regardless. Technically only the UTXO set... past a certain window count of confirmations sufficient to discourage all future spending / mining attacks upon any shorter windows of it... is relavant to most BTC style and other blockchains. The window size is some estimate of excess of future computing power and other attack methods and surfaces, converted to a block height. For example, one year to six months of confirmations might be sufficiently immune, perhaps even approaching applied real world difficulty of compromising a / many signing key[s] eg: UTXO's. History beyond that window count is irrelavant to determining real world validity of a UTXO. History's only use thereafter is for anti-privacy dataminers trying to regulate control blackmail censor tax terminate etc, both you and cryptocurrency. These anti's are secretly embedded in every cryptocurrency project, and abound in the rest of the world, and will resist proposals to create history free coins and operation. They'll datamine anyway the stupid non-private coins like BTC, but they can't do that with strong cryptographic privacy coins, like ZEC, that use ZKP's. Users simply do not need that history on disk, nor do nodes need to be passing it around as bootstrap. Users can set a flag to listen for and retain their own histories if they need some blockchain style proofs for whatever purposes. And just as users might pay for their own disk / bandwidth to maintain history since Genesis, they can just as easily pay the dataminers instead, up until even the dataminers possibly cannot afford to keep it, or the irrelavance of age hits. To do this the software upgrades to a protocol that not only mines the UTXO set into a form of current meta transaction blocks like BTCP did to bootstrap ZCL + BTC = BTCP into its launch... but also does this checkpointing periodically at programmed heights, adding confirmations to the latest UTXO meta tx block sets. Everyone holds at least one sufficiently confirmed set which is then used as bootstrap for clients similar to BTCP. Clients refer to the recent set and the confs over it to validate received value. Clients will scan forward from a well confirmed set to see if something was subsequently spent in the usual fashion, thus avoiding race and poorly confirmed status of newly published sets. They might also consider the embedded signatures of some plurality of miners over UTXO sets for bootstrap purposes, or some live network method, including referring to the dataminers, to serve any need to fill in for now missing genesis path. Thus the impact of natural tx growth on storage space and validity referencing CPU time becomes smaller, now expressed in terms of size of the few UTXO sets needed (perhaps no more than two or three) and their window sizes. That's one of the ideas that has been out there for existing blockchains. Another one is splitting and distributing the data storage for reference across the p2p nodes via DHT indexes. ie: 100000 nodes, 10 levels redundant... a million nodes, smaller than even the bittorrent cloud. The ideas seem more like underhood scaling optimizations, than being "sidechains / lightning / forks / alt / shit / central" solutions (as some say now affect BTC) away from further exploring closer to original "Satoshi' architechtures. A lot of people are now looking closely at BCH with its rising adoption (even now surpassing BTC in areas), and others, to be shining examples of that exploration. Who knows, it's still early days for all things cryptocurrency. The right combination of factors leading to long term adoption of say a max of five leading coins covering the range of blockchain use cases probably hasn't been found yet. Stay nimble my friends.