Rahko and Bleximo. They focus on quantum machine learning, and ‘practical’ quantum computers for quantum simulation, respectively.
The big bet for Rahko and the rest of the wildly proliferating quantum software startups must be that some sort of useful, usable quantum computer is going to exist in the near term (< 5 years). If it doesn’t, what they will be left with is a smattering of arXiv publications and some preliminary validation of their methods on whatever toy QCs IBM, Google, and Rigetti choose to provide for public use.
Bleximo’s case is more interesting. It’s clearly a company founded by experimental physicists, so they must intimately understand the current and probable future limitations of the technology. Although their website is spartan, more information is available here. Their approach appears to be to develop near-term QC as a quantum-coprocessor rather than a stand-alone computer for simulation tasks rather than general purpose computing. It’s not clear to me that the current state of the field will support even that application, but as the DoE badge shows, there is interest here that may survive the inevitable quantum winter and provide more reliable funding throughout the dark years ahead.