All HFTs I'm aware of have a continuum along developer_speed to execution_latency. On the developer_speed end we have things like Python (or Perl, Ruby, etc) ->

All HFTs I'm aware of have a continuum along developer_speed to execution_latency. On the developer_speed end we have things like Python (or Perl, Ruby, etc) -> Java/C# -> C++/C/assembly -> programmable nics/fpgas/gpus with access to the nic -> custom hardware. At each part of the path development gets much harder, dev times go up, talent gets harder to find, and the whole thing becomes much more expensive. The trick is use as little of the bottom as you can get away with and control it from as high level as you can. It would surprise me more to find an HFT that doesn't use Java at some part of their execution lifecycle. Not everything needs to be in microseconds, nanoseconds, or picoseconds, especially if you can factor it out of the critical path and do the work/build some data structures ahead of time.