I think a big part for them is to prove their AR software. They will also need to a have killer app (game demo) created, proving (POC = Prove of concept) rather that a video sample. What they want to accomplish is cool indeed. I lead a software development team and this sort of things can be tricky if not done right. Unlike the Rift, the games can be fully controlled as it is 100 percent created by an engine that has predictive parameters. AR is mix media and although it can be done, it brings a whole set of challenges that are not typically presented in VR. The Zombie sample video is cool, but as games need become more dynamic and user movement is required, their AR software algorithms requirements to interact with the real time footage will prove more technically complex (my opinion).
At the end this is all an opinion, I feel they will need to have a real-time running demo and be a well thought out experience. Ultimately, the experience must be unique one if to take the focus away from the Rift to a degree. If you are going to be ambitious then is great if you have killer demos that proves what you sell as a way to silence any doubts.
One area, that I applaud the Rift is that they have been very clear and focused from the beginning. They have been focused on games and making the experience the best they can. Certainly, having the voice of Carmack, the experience of Iribe with deep industry connections is a big help. Naturally, from games the technology will grow to other segments of industry organically.
On the hardware side, I am intrigued. I wander if they will attempt to use Fresnel Lenses. I am not sure you if they can attain 160 FOV from a typical aspheric lens. Maybe they use bigger lenses compared to the Rifts and with more augmentation.
This may sound like I am knocking their prospects, however I am stating some opinions and questions.
But all in all it sounds great in theory!
this is what I was suggesting to oculus from many days back 
