There's a difference between a full node and an archive node. A full node has all the data required to verify the chain from genesis up to the current state. An archive node also has all the state of all previous blocks. An archive node can be derived from a full node, and the only purpose of an archive node is to run historical queries.
You can run a full node without staking ETH, in which case there are no penalties for going offline. (And if you are staking, you can be offline up to a third of the time and still be profitable.)
There are efforts underway to shrink storage requirements, including very secure light clients, state expiration, and statelessness. But right now that stuff is mostly on the back burner while they focus on the proof-of-stake migration.
https://etherscan.io/chartsync/chaindefault
There's a difference between a full node and an archive node. A full node has all the data required to verify the chain from genesis up to the current state. An archive node also has all the state of all previous blocks. An archive node can be derived from a full node, and the only purpose of an archive node is to run historical queries.
You can run a full node without staking ETH, in which case there are no penalties for going offline. (And if you are staking, you can be offline up to a third of the time and still be profitable.)
There are efforts underway to shrink storage requirements, including very secure light clients, state expiration, and statelessness. But right now that stuff is mostly on the back burner while they focus on the proof-of-stake migration.