weak Cards though.. I ordered a 1080, has twice the power... and for VR you Need all the power you can get... 2 of those 480s is also no Option as currently neither crossfire nor SLI works properly for VR
Not necessarily: The RX 480 has the same GFlops power as the R9 290X. If we remember, this was the card that trounced the Titan for about half the price.
I recently upgraded my graphics card, but the 290X scores a 6.8-7 on the VR Performance test. Before I upgraded I could run VR High with most settings set on high, with the exception that I couldn't use Supersampling.
The RX 480X has the same amount of power for $199, so it will make it possible for those that cannot spend 500+ on a card to get into VR.
Here's the thing: Nvidia is coming out with this "multiprojection" and all that stuff that supposedly will make VR better. It is no secret that AMD is quite ahead of Nvidia in a few aspects.
For example: Nvidia's DX12 performance is sad, to say the least. Even though we do not use DX12 much now, it's only a matter of time before DX12 slowly replaces DX11, just like DX11 replaced DX10 and 9 before it. For all the hype on the GTX 1080, it's only about 6% faster than a Fury X in DX12 performance. The Fury X was in turn a good amount faster than the 980Ti/Titan. This does not bode well, since when AMD pulls out their own FinFet cards, they will trounce Nvidia on DX12 by a large margin.
Asynchronous Compute Shaders: AMD has supported Asynchronous Shaders for quite a while now. My 290X card, despite being about 3 years old, was already able to take advantage of the asynchronous shaders (which is a DX12 feature. The 290X, while being an DX 11.1 card, was actually able to run some DX12 features despite its age) Nvidia does not have this capability. Or rather, they haven't bothered to push for it. Asynchronous shaders are a great asset for VR, and is one of the key features in DX12 that will make VR much easier to run.
All-in-all, despite the outrageous claims that the "1080 will be twice as fast as the Titan", when in reality it's about 20-25%, the GTX 1080 is a huge letdown. I don't know how this "multiprojection" will help VR, but Nvidia is basically playing the catch-up game in the VR field. AMD has been working hard on VR performance and support, as shown, by the fact that an old R9 290X can perform extremely well even in a game like Elite.
I actually sold my R9 290X and bought a 980Ti not long ago. Asides from being able to use Supersampling (1.5X, 2.0X actually drops the FPS to 45 in rings systems, which is what the Vive locks the FPS to if you can't keep up 90FPS), there isn't much of a difference in VR. A 3 year old card *almost* keeping up with a 980Ti in VR, (again, except for Supersampling) that's not a good sign. Let's hope this "multiprojection" is the solution to this problem for Nvidia.
Polaris 10 and Polaris 11 are coming this year. Vega is coming next year. The RX480 seems to be the lowest end AMD card on the new lineup. And it pushes 5.6TFlops of performance for $199 (GTX 1080 pushes 9TFlops if I remember right). This means that very likely, the RX 490 might perform equally, or faster than the 1080 (adding in, of course, the better technology support necessary for VR).
Explanation about asynchronous shaders and why they matter to VR: In a typical scenario, such as DX11 and before, the shaders on a card perform one task at a time. Yes, they do it extremely fast, but they still have to process information in a pipeline. Even though video cards today are extremely advanced, and have thousands of these shaders, they were tied by the API. Asynchronous shaders is a feature first seen in AMDs Mantle (let's be thankful, if AMD hadn't developed Mantle, DX12 wouldn't exist. Mantle pushed for a whole new level of advancement in graphics API, such as DX12 and Vulkan).
Basically, Aynchronous Shaders allows the card to use all its shaders (you guessed it) asynchronized. In simpler terms, it's multi-threading. Shaders are no longer forced to process information in a pipeline, but instead, shaders that finish a task can immediately pick up another task to speed up the process. The issue with Nvidia at this time is that Nvidia cards are unable to do this at the hardware level, although they might be able to force it through drivers. All AMD cards since the R9 series support asynchronous shaders at the hardware level, and have been designed with the feature in mind.
You could think of it as Oculus and Vive. Yes, the Oculus can do room tracking, but it will never be as good as the Vive that was designed from the ground up for room tracking. Nvidia will be able to do asynchronous shaders, but unless they design a card with the feature in mind from the ground up, it'll be the same scenario: It won't be or perform as good as an AMD card. (When this feature is concerned).
How does asynchronous shaders benefit us? It could make Crossfire perfectly viable for VR. For example, in DX12 cards no longer mirror their VRAM (you know, normally having 2 4GB cards means you still only have 4GB of video memory). In DX12, the cards memory is pooled, so you effectively do get 8GB of usable video memory when you put 2 4GB video cards together. By extension, this means that in the specific case of VR:
You have 8GB of video memory, 4GB per card.
Due to the asynchronous nature of the feature, cards can run asynchronously.
In simpler terms: You could have Card # 1 render Eye 1 on VR, and Card #2 Render Eye #2 on VR. This could mean that even a low end card like the RX 480 would only have to worry about rendering a single eye. Something that it definitely has more than enough horsepower to do.
This is not a fanboy post, by the way

I'ma computer engineer, and I simply buy whatever offers the best performance. This also means doing extensive research on features, and what each hardware can or cannot do. I own a 980Ti, owned a R9 290X before that, and had SLI 680 before that. When the time comes, IF Nvidia is the better choice, I will buy that.
Doesn't look like that will be the case, though.