Between AR and VR, AR is the harder to address comprehensively - we knew going in that we can make pretty pictures in VR by leveraging existing CG pipelines, and there's definitely some market for aesthetic immersion in VR. But VR has also mostly hit a wall outside of that realm since it gives up a lot to achieve this goal - cost and accessibility are both much weaker relative to 2D screen experiences. It's badly positioned for consumer use despite intensive marketing in that direction. AR adds the extra claim of making use of more of the real world, and this is still a thing with potential, but again has trouble in many cases justifying itself relative to the headaches - for a lot of applications the quality and kinds of sensor data needs to be higher for it to really give results.
Net result: a bunch of startups got funding to chase after the buzzword, the survivors mostly pivoted into more traditional businesses.
I guess this market is waiting for wearable AR hardware. Oculus should release consumer product in next 12 months, and Apple is also cooking something big behind curtains.
On the other hand, I could also see the tech just fizzle out completely because it'll turn out that most of the population will get motion sickness from VR experiences regardless of tech advancements.
Heck, besides Angry Birds, very few transitioned well to touchscreens. Touchscreen only works because it's a convenient platform, not that it's better at anything.
I think it needs some creative effort as a lot of it is still groundbreaking, and people have difficulty thinking in the "real world" dimension. Remember, a lot of our current technology was probably influenced from sci-fi like Star Trek first. It took some inspiration for operating systems to go from pure text based to windowed, and some more time for users to accept it.
It's possible the first practical use of AR might be like the Iron Man scene where Stark uses it to view the suit in 3D and manipulate the controls. It seems like Magic Leap is working on something like that for education.
I have this feeling that there's a few really good, but not flashy uses for "AR", things like measuring angles and lengths at a glance in carpentry and construction. These are obvious but not pursued because the "Virtual Reality Will Make Billions" scam / reality distortion effect stomps all over them, and no idea that's not worth billions will be discussed.
Also, "VR/AR" displays are probably progressing, (surely the horking problem has been solved by now, right?) but where's an input method even half as good as a keyboard?
We're in the interstitial space between the technology becoming available and widespread, and what will foster that adoption is for someone to make the first great AR game.