My guess is they measure pretty well with perhaps some intentional deviation from Harman but not much. Distortion will come out OK but not amazing. Group delay might come out a bit of a mess due to all the DSP. ANC performance (AFAIK) is not something Amir measures, but I expect it to be within 3dB of Sony's.
Sound-wise, what interests me most is firstly whether or not the feedback system they're using will be robust enough to ensure a consistent response at low frequencies across individuals and positions. I think they use a Qualcomm chip inside, and I have read nothing about them using other adaptive / active systems developed on their own, so my guess is that we'll get a fairly standard feedback system effectively operating up to around 500-800Hz. In which case I hope that they've designed the Ace to minimise positional and inter-individual variation above the range where the feedback system operates. And then (and then only) whether or not they're using a sensible target to begin with

.
Excess group delay (what really matters) should be fine, as it is for most (but not all) headphones. The background noise from the active systems could raise the noise floor high enough that it will affect regular group delay measurements, but you shouldn't rely on them to tell you much about the headphones in the first place anyway - and if that is the case then the noise floor itself would be a significant detriment to sound quality (ex Bose QC45 among others) !
I think the real reason to buy these would be the spatialization effects they're working on, with the head tracking, etc. That could actually be very cool. Not sure if I would throw $450 at wireless headphones at this point in my audio purchasing journey (already have more cans than I know what to do with) but for a late night home theater-type experience they might be cool if you can make use of the spatial effects.
It's their own renderer so it will be interesting to see whether or not their generic profile works well, but I am not certain that it's necessarily going to be better than other solutions already there. Besides we're now moving into at least partly individualised profiles, so...
What I think is the most interesting with their soundbar + HPs solution rather is the promise of being able to use any input (console, streaming box, the TV itself) and send a binauralised content to the headphones with presumably low latency. I am not sure of this but I don't think that other solutions seen so far (ex Apple TV + AirPods) are quite as all-encompassing in terms of inputs for example.
The "simulate the room you're in" feature should be pretty interesting. On one hand, should be very impressive in terms of realism.
I am not certain of this. The only thing they can realistically do is to mimic your own room at low frequencies for the most part. The issue at higher frequencies is that they'll still need to use a generic HRTF map instead of your own, and since the headphones will cover your ears when performing these measurements, they won't capture anything related to your own anatomy (other than maybe have some idea about the width of your own head).
On the other hand, my room ain't all that. How about simulating a good room instead?
Without turning that feature on, what I understand is that they binauralise the content in a virtual room that they consider "ideal" :
I also would probably rather go with the default, "ideal" room than my own right now !