The even more impressive here is the power/weight ratio - about 2000hp/kg (whole Buggati engine in 1 kg). (My speculative estimate - the Raptor turbine in the turbopump can hardly weight more than 500 kg as the whole engine is 1600kg -> whole turbopump less than 1000kg -> the driving turbine is half of it max). So anybody who plans to beat Musk/SpaceX (like i do when i finally get a real garage :) has a very high bar to clear.
Additional facet is that Raptor turbopump is full flow and thus runs at very low (relative to other gas turbine machinery out there like for example F-16 at 1200C+) temps like 500-600C which means that the power can still be almost doubled with the same regular materials they use - steel and Inconel (and this is what we're seeing - about 1.5X power increase from Raptor 1 to Raptor 3 in a span of mere few years while the engine weight is even decreasing a bit)
But even MORE impressive is the current raptor engines cost between $1m to $2m each with a target cost of $500k when in full production. The F-1 cost an inflation adjusted $85m-$100m each.
> Additional facet is that Raptor turbopump is full flow and thus runs at very low
You write it further yourself, but rocket turbopump turbines aren't running with cold gas - the hotter the input gas, the more energy can be provided to pumps, so turbine blades have roughly the same requirements as those in F-16. The design on those blades is pretty complex, with materials, processes and substructures like internal cooling channels - all to reach possibly higher temperature to work with.
Additional facet is that Raptor turbopump is full flow and thus runs at very low (relative to other gas turbine machinery out there like for example F-16 at 1200C+) temps like 500-600C which means that the power can still be almost doubled with the same regular materials they use - steel and Inconel (and this is what we're seeing - about 1.5X power increase from Raptor 1 to Raptor 3 in a span of mere few years while the engine weight is even decreasing a bit)