AMD may beat them to the punch.
Also, why would you say that hololens doesn't count?
It doesn't do what, for me and many others, is the defining trait of Augmented reality - it doesn't occlude computer generated objects around reality. By which I mean it needs a method of derviving not only the geometry but also spacial mappings surrounding the headset quick enough and reliably enough, and be able to derive it's position in the room, to make objects appear to pass behind real life objects in front of them.
Hololens doesn't do this yet. It can detect surfaces and make it's projections look like they appear on those surfaces, but if anything passes in front of them the illusion is broken. It's not mapping the environment, nor is it discerning it's position in the environment. This, to me, puts hololens in the realm of Heads Up Displays, which are a precursor to AR (along with VR).
Ultimately, the companies which will likely get us environment tracking that is fast and accurate enough for AR use will likely either be google or microsoft, as they've invested heavily in R&D projects designed to do that. But neither Google Glass nor Hololens are what I'd consider AR yet. They're just very forward thinking HUDs.
You can already see the sectors of computing that will need to converge to make AR a thing. We'll need environment tracking to get great, we'll need Oculus and Valve to continue solving the logistical problems associated with Virtual application design, and we'll need Nvidia, Apple, and Oculus/Carmack working to make mobile computing viable for powering this hardware in a small enough form factor to be self contained.
That's why people following VR and AR agree that AR is still over a decade off. Too much needs to happen before it becomes really viable, but the seeds are already being planted.