It seems that NVIDIA does not want to miss that opportunity and will be pushing to get Pascal with HBM 2.0 on the market around the same time. This is the current rumor floating out of NVIDIA although we are not 100% sure that this holds water. HBM 2.0 is a very attractive option that both AMD and NVIDIA want to get out to gamers, but the problem is still one of supply. From everything we are hearing HBM 2.0 is not at the levels it needs to be to meet the demands of either AMD or NVIDIA. This means that a Pascal GPU from NVIDIA would either be a paper launch with some media and partner sampling or it would feature a different memory standard.
One possibility is that NVIDIA could use GDDR5X in the early cards. This specifications for this have recently been published, but we have also heard rumblings that NVIDIA might be leaning on it for some early release cards. This would also roll into mid-range cards in 2H 2016 and 2017. GDDR5X has roughly twice the bandwidth of GDDR5 and is a great stop-gap until HBM is more readily available.
Of course, it would not be the first time that NVIDIA (or AMD for that matter) did a paper launch with just enough cards to hit the media and partners. So we could see Pacsal with HBM 2.0 “launch” in April, but not hit general availability volume until late 2H 2016 or 2017. We will have to reach out to a few sources to see if we can find out the exact status of HBM 2.0 or if NVIDIA is trying to get production ramped up to meet a launch deadline. Either of these could be supporting evidence that Pascal could hit the market around the rumored time.
It looks like VR will be the factor that pushes both NVIDIA and AMD to churn out better GPUs for this year and leading into 2017. This is good news for consumers and the PC market in general if both companies can make enough to keep the on the shelves at a decent price.