Given enough time, lost coins will be reclaimed by technology that is able to claim them. Of course when the time comes non-lost coins will not be vulnerable to the same technology that claims lost coins.
This is indeed a possibility for those old miner coins nobody is spending (many of them are supposed to be Satoshi's) - think of quantum computers able to break ECDSA. The P2PK transactions could be "cracked" in theory.
Just recently there was a discussion about the economic consequences of this scenario -- let's say quantum computers (or any other future computing technology) gets strong enough to break the keys because P2PK transactions expose the public keys, and Shor's algorithm could in theory calculate the private key taking the public key.
Once this becomes a possibility, opcodes for post quantum cryptography could be added and users be heavily recommended to transact their Bitcoins. And those who do not, have used P2PK and have re-used their address, would be vulnerable to this attack.
Altcoins have often already performed similar swaps, so it's no nuclear science. The only problem is that if there's much onchain activity and the change comes too late, then there could be a huge fee spike, so it would make sense to add post-quantum cryptography challenges relatively early.
And once these "presumed lost coins" which don't participate in this "swap" (not the "provably lost coins" like those burnt in OP_RETURN transactions!) are attacked, then they would be redistributed by the hackers (I don't believe they'll try to HODL).
Maybe even today with
BitVM you could already create a post-quantum challenge? This would actually an interesting idea (Edit: I opened a
thread about this for those interested).
The "quantum computer attack" is however a
very remote possibility and also has been discussed several times here.