Due to the large performance benefits, we early upgraded all of Uber's mobile engineers to the M1 Max (64gb) a few years ago which gave 2x the build perf vs previous gen intel.

Some of those folks are coming up for refreshers, so we did a new analysis vs the M3 Max (64gb) and we're seeing about 35% reduction in build times vs the previous cohort.

Screenshots from our benchmark harness which tests clean builds and several different build graph invalidations to mimic common dev flows.