I mean, why Bitcoin nodes keep all the data of all blocks, back to the genesis block? The majority of this information is not required for processing new blocks and may be ignored. Well, some statistical resources should still keep it, but to run a node it's possible to just delete the majority of historical data. Am I right?
Once a database of the UTXOs is constructed, the blocks are no longer needed. The blocks are kept in order to provide them to other nodes and to reconstruct the UTXO database when necessary. Blocks must also be kept for resolving branches.