Near as I can tell, it's the only form of practical quantum computation that's available to end-users in any meaningful way. Everything else so far is vaporware with ridiculously high error rates.
> (it is not particularly "quantum" either)
The qubits - josephson junctions and superconducting loops - are indeed in quantum superposition with each other if their claims are true. This is a macroscopic quantum-mechanical effect; to call it "not particularly quantum" belies a lack of understanding as to what they're doing.
> most researchers are fairly annoyed at D-wave for originally pushing this nonsense
D-Wave is making their system publicly available for free. Have you signed up and taken a look at their demos? Admittedly it's a bit beyond my mathematical ability but it's clear the basic seed is there for something that works; they just need a bigger graph and more connections between qubits. There's no reason to assume they won't have their own equivalent of Moore's Law at some point where these things grow year over year.
> transmon qubits at Google/IBM/Yale for some well known prototypes of quantum computing hardware (which are admittedly quite far from being practical or usable).
Do they have _any_ results worth caring about yet? I don't think they're anywhere near something working on a practical scale. You can buy a D-Wave machine, or time on it - it's not vaporware, even if most people don't find it useful. For the first few years of gate model machines people aren't really going to find them all that useful for day to day things either - just like computers in the 1940s and 50s, it was very specific, high-dollar applications first, and things for the common man decades later.