On the contrary, compressed blocks are smaller to transmit over the network, so it's faster. The block is compressed by the block producing node (witness) and it takes a very tiny fraction of CPU processing. It's well explained in paragraphs 2-3 of this post.