I never see the Tezos node using more than CPU core even when syncing from scratch. Why is this? Is this the reason why the nodes takes forever to sync in archive mode? Are there any plans to make the node multi-threaded or more performant?
The performance of syncing an archive node depends on multiple factors including your hardware, network connectivity, number peers etc however one of the dominant one is the intensity of I/O operations not the number of computations by the CPU (see this question for instance). It is more likely that you would go faster by improving your storage from say hdd to ssd than by having the node go from single threaded to multithreaded.
To fasten up syncing there are several alternatives