No one ever seems to talk about this but sometimes... reviewers are terrible.
Not terrible like "Ugh they made me revise my paper" but terrible like "Did this person even read the paper?" I once had a paper where a reviewer made lengthy complaints about the terrible user studies and lack of clarity in user studies for an object. There were no user studies in our paper, there were several clear statements about how there were no user studies in the paper, the 'future work' section detailed the user study collaborations. It was a paper on the method of integrating physical and automated controls, not a paper on user studies. One of the reviewers just... lost the plot. I don't know if they read a different paper, I don't know if they wanted a different paper. The other two reviewers made some excellent revision suggestions which we followed.
It wasn't the first time either. I can't speak for all fields or conferences, of course, but sometimes you just get a terrible review. Terrible not negative, terrible as in makes no sense. Sometimes you respond, sometimes you try to work it in and sometimes you just politely ignore the bits that make no sense. Some fields lend themselves to a wide variety of subject matter experts. Consider, for a moment, robotics. A submission to a robotics conference or journal could be about the mechanical engineering aspect, the automation(AI and Machine Learning) aspects, optimization, computer vision, sensor integration... No single person is going to be an expert in all of these things and some of the most interesting, to me, projects involve some blending of these subjects. It can be hard, then, to find a reviewer from the pool who best fits a submission.
This is the case in many subfields in Computer Science, this may be the case in the field you are working in. When you're talking about research on the bleeding edge sometimes you're talking about stuff that has a limited or restricted audience and, thus, a limited or restricted reviewer pool. There are a fair amount of similar questions about academic research reviewers and reviews here. Typically the benefit of the doubt is given that the reviewer is bringing up legitimate complaints and, to be honest, that is the best policy for everyone to have. But sometimes it's ok to accept that a reviewer is totally out there and the review is unhelpful or inappropriate(in that it does not apply.) In those cases I would respond to the other reviews as appropriate and respond, tactfully, to the review in question with statements directly reflecting how the review does not apply to your work.