If you watch the first stage telemetry as it comes back down on the livestream (they don't always show you it, sadly) you can see that actually a huge proportion of the velocity is lost through atmospheric drag. I was shocked the first time I saw it.

Obviously the terminal velocity of the stage is still pretty fast but compared to the speed it starts at it's much smaller.

Yea, atmospheric drag does most of the work to slow the boostet rocket. Which makes it all the more counterintuitive to my brain that the economics and engineering doesn't make sense for deploying some kind of chutes to shave off velocity.

Every kg in fuel you need to burn to land, you need to lift off the pad. From what I found through some Googling, a F9 booster uses a little under 1,000kgs to do the hover burn. I would imagine some lightweight aero surfaces would be more weight-efficient at decelerating the rocket than the 999th, 997th, 996th, (nth) kg of fuel.