So you are one of those that believe that the 380X will be as fast as the 980? Then what about the 390X?

@Sentient_Nebula: What they heck are you talking about? The 295X2 smokes the 980 is most situations and it's been available for a lot less than $1,000 for a long time. And with its closed liquid cooling it's also cooler than a 980. But you're absolutely right when it comes to power consumption.

The 295X2 is far beyond the call of duty (pun intended) and it is a waste if you are not running ultra HD Monitor that can provide the res this card can crank out so if money isn't the issue, first consider what it is you are trying to achieve and what you will be viewing it on.

Also, and especially for AMD and the 290x and 295X2 be sure not to skimp on quality PSU and shoot for Gold cert at the minimum to max power efficiency.

Finally, as counterintuitive as it seems, select a case that fits the airflow, additional cooling and space you will require for the parts inside. Personal I adore Cooler Master but you have tons of choices.

Happy Rigging!

AMD's yet to update the 290 series. But the 290X performs about 10 fps under 980 usually. They'd both crush any 1080p game so either way you go you can't go wrong. Although AMD's 390X will probably be on par to maybe better than GTX 990.

AMD's cards are usually cheaper, more power hungry but perform better with some games. The 295x2 beats the Titan Z every time, and the Titan Z costs twice the amount of a 295x2.

I don't think AMD will ever release an R9 390X cause the yields for TSMC's 20nm process node isn't up to par yet to handle a ~400-500mm^2 die. I think there will an R9 490X by the time TSMC ramps up their 16nm FinFET node, the initial yields should be better than their previous 20nm node due to the fact that they both have identical transistor gate pitch so they should know the quirks better in the end and that should give AMD enough time to do an update to the GCN architecture up to 2.1 ... 

I think the R9 380X should be FASTER than the GTX 980 and I also don't believe in the GM200 or GM210 hype either ... 

GK110 was only able to get so many gains against GK104 because of the fact that the former had a much larger die space to work compared to the latter whereas GM204 already has a substantially large die ... 

All focus on GTX titan.

LOL. Yeah, I'm not thinking too clearly tonight. Had a long week. I usually don't like to edit my posts though, as I feel that can cause me to try to deny things I said. (I like to keep the truth in place). I usually just use edits to fix spelling or grammer errors.

I checked a benchmark to verify what I said just then, but looking at it again, I noticed they were running a 295X2 vs SLI 780tis (With the 780ti being similar in performance to the 980). So yeah, a 295X2 is similar in performance to SLI 980s, not just a single 980.

If you want an equivalent in the same performance class, next year we will see one. But don't count on it being as power efficient as the 970/980.

It could be...

By the way, TSMC doesn't mention AMD in the short list of 16nm FinFET adopters and that, paired with the fact that GloFo and Samsung have a partnership for a 14nm process has lead to some speculation.

Since their split, AMD has had a contract with GloFo to buy "X" (I don't know how many) waffers each year, but since almost the begining AMD hasn't reached that level of demand so they have been paying them a compensation for basically nothing while buying from TSMC. And now GloFo will have a more advanced process than TSMC with the 14nm but also with the 20nm one as they have access to Samsung's expertise on that node. Get both things together, add the fact that TSMC hasn't delivered on their schedules for the last node jumps, and you can see how AMD could skip TSMC althogether from their 20 and 14nm products and go back to GloFo.

But of course, it's just speculation.

@Sentient_Nebula: No worries.

First off, it's just a short list of partners ...

Second, it's possible that AMD are looking towards Global Foundries to manufacture their future graphics chips. 

Third, I doubt that either Samsung or Global Foundries are developing a 20nm node so they'll probably have to look at TSMC for the time being until the pair can ramp up their 14nm process node for mass production. They just skipped 20nm altogether and it will be short lived on TSMC's part too. 

1-Yes, it's a list of their 16nm FinFET process partners, but it's "curious" to see Nvidia on that list but not AMD.

2-Nothing else to say.

3-Samsung already manufactures nand/other chips on their own facilities using a 20nm process, they don't need to develop it. That's why I said GloFo could use their experience with that node. And AMD is looking forward that 20nm process for the XboxOne and PS4 APUs as noted by a beyond3D posts picked by Eurogamer.