Not a great comparison, the 10 gig switch is mostly SFP ports which are only useful for short run twinax or with fiber optic SFP modules for anything beyond twinax lengths. 10g copper SFP modules don't exist. Useful in a rack with servers with SFP NICs or if you want to fuck around with fiber, but in my mind that rates them as less useful than base-T which has much simpler and cheaper cabling demands.
I see a lot of twinax/optical deployments as converged core server + iSCSI storage but mostly in new cluster deployments where the expectation is everything is new and there's a few fiber handoffs or for core network deployments in larger networks.
But the most useful is always the base-T version because it drops in easily and handles pre-existing equipment with only 1g copper connections.
To be slightly fair with switch vendors, there is something complex about 10g-baset PHYs which makes them more expensive, but not THIS expensive for this long.
I still think IEEE messed up by not rolling variable (2.5/5/10) link speed into the 10g-base-t standard up front. It would have driven switches with broader footprints and driven more adoption by giving full speed where the cabling was good and 2-5x speed where cabling was just OK. More adoption, more unit volume and lower prices.