Last year everyone was expecting that Zen will be the product that will turn around AMD. Only Zen.
Latter it was becoming obvious that DX12 could also give a chance to GCN architecture to be more competitive, and who knows, even gain an advantage against Nvidia's architectures.
If VR succeeds, both AMD and Nvidia will gain, but the real story here is that AMD's designs, compared to Maxwell at least, are not inferior to Nvidia's for use in VR.
Then we have Intel's problems to maintain a good lead over their competitors in the manufacturing market. That's good news for AMD in the x86 market considering that they could be fighting Intel's 10nm chips, with "10nm" manufactured chips from TSMC of GlobalFoundries in 2018.
On all those, now we add three new semi custom designs that warranty AMD's income from consoles, and this latest joint venture with the Chinese. This last deal with the Chinese, not only open the doors to a market totally lost for AMD those last years, it's also an indication that Zen is nothing like Bulldozer. If the Chinese wanted just an x86 chip, they could go to VIA and save that company from nothingness giving them peanuts, or they could have done this deal with AMD months ago, or even years ago. But I guess they where not that much interested about Bulldozer architecture.
Of course 90% of the above are still speculations and best case scenario. Things could turn around to the worst very easily. For example, Zen could be still too little too late(it will not be the first time) and Pascal could kill Polaris in the gpu market(Nvidia had all the money and time in the world to create a beast). US and Intel could try to block this joint venture and there is always the possibility those three semi custom designs to not be all for consoles. A new Nintendo or Microsoft console that doesn't use AMD hardware could be a big shock for investors and sent the share price the other way.
