Quoting a response to a similar question:
Reducing block time is a priority, but we want to be confident that we can do this safely on Mainnet. We have ongoing projects, both on the Tezos protocol side – like the validation pipelining work that has (partly) been rolled with Kathmandu – and other ongoing work on the Octez node, which target to reduce block validation and propagation time. These are a prerequisite to reducing block time safely, otherwise we risk having cascading effects with higher congestion and slower blocks.
