AMD confirmed they won't be releasing the HD8000 this year anymore so nVidia is also likely to introduce the GK110 GTX in Q4 or early next year aswel.

Its better this way, they will have nearly 2 years development cycle and thus will provide a huge upgrade from GTX600/HD7000 series.

Both GTX780 and HD8970 will have more than 7bn transistors and around 6TFLOPs. The difference with current enthusiast cards will be HUGE.

zero129 said: I would buy one if i had a need for it. But with the way Nextgen gaming is looking i think i will get away with a much lesser card and still be able to do 1080P@60FPS maxed . Im just waiting for a game to make me need to upgrade again

Would that be when DirectX 12, and the next OpenGL, come out? 

Anyone know what kind of hardware they are waiting for that?  Perhapse Windows 9?

Zappykins said: zero129 said: I would buy one if i had a need for it. But with the way Nextgen gaming is looking i think i will get away with a much lesser card and still be able to do 1080P@60FPS maxed . Im just waiting for a game to make me need to upgrade again Would that be when DirectX 12, and the next OpenGL, come out?  Anyone know what kind of hardware they are waiting for that?  Perhapse Windows 9?

Kinda like how they trojan horsed vista for 10 and win7 for 11? :(

Zappykins said: zero129 said: I would buy one if i had a need for it. But with the way Nextgen gaming is looking i think i will get away with a much lesser card and still be able to do 1080P@60FPS maxed . Im just waiting for a game to make me need to upgrade again Would that be when DirectX 12, and the next OpenGL, come out?  Anyone know what kind of hardware they are waiting for that?  Perhapse Windows 9?

Rumour is that DX12 is in development hell for non-technical reasons.

But honestly we don't need DX any more, it's programmable enough so that any feature can be implemented via the shader compiler.

Windows "9" doesn't exist, it will be a yearly release thing after now.

Huge, but expected. Probably as much as the 280 > 480 transition, something around 70-80% of performance increase and the usual for 18 months of wait.

However instead of remaining around the same TDP, their power consumption and dissipated heat is going to scale up with the extra power. A full refresh would be expected a year and a half later, but they are delivering about the same thing the 500 series was (while it released a few months after the 400 series, by the way) because the TSMC manufacturing processes can't keep up.

Basically they are faking a real evolution because they have to launch something.

So it's going to be around $1000? Damn, that's pretty expensive. It's the same price as the gtx 690, but you get less performance, albeit in a single card package.

I just sold my GTX 580's and I think I'll probably be picking up new card(s) soon.

From CGI's link above:

So the launch date is 18th February, it is being said that only a paper launch would occur with retail availability around 24-26th February. The GeForce Titan would be available in limited quantities as expected from a high-end flagship card and would be limited to certain manufacturer’s such as EVGA, ASUS, MSI. Pricing is said to be around the $1000 mark but you could never be too sure about NVIDIA’s pricing scheme.

How does a team of CPU engineers fleeing to Apple have anything to do with AMD's GPUs? There are many other reasons why AMD has no interest in making a $900 Titan competitor and never even intended to so do:

1) AMD doesn't have a strong Professional Graphics unit that sells GPUs to corporations (Tesla and Quadro), but NV does. For that reason, all the failed GK110 chips that couldn't be sold as K20X chips can be resold instead of just being thrown out. *Hint: the Titan will not have full double precision performance required for the Tesla markets due to issues with the chip. Watch, the double precision computational performance on the Titan will be completely crippled compared to K20X parts because NV will use salvaged K20Xs chips. Then they'll take advantage of the additional power consumption headroom left by the "turned off/broken" DP compute transistors to clock the actual GPU higher* Becase NV shares its R&D for GeForce and Tesla/Quadro product lines, it allows them them to build 500-550mm² die chips. AMD cannot afford this luxury and never did if you revisit all their GPUs back to 2900XT in 2007. To suggest AMD ran out of $ to compete with NV in the large die strategy ignored completely what AMD has been doing since HD2900XT series. Without the Tesla/Quadro products' success, the large die 500-500mm²die strategy would likely not be possible.

2) AMD has never built a 500mm² die GPU chip and neither has ATI. For that reason alone, once NV continued with its big die strategy, NV having a class leading GPU on the high-end was a foregone conclusion since GeForce 8800GTX in 2006. That's 7 years ago. Your comment implies this just happened. Nope. AMD was never competitive with NV's flagship ~500mm² chips.

3) AMD is focusing its efforts on HSA and compute. GCN was designed to be a well-rounded GPU for both of these purposes, not only games. The Titan will be a dog for OpenCL and DirectCompute. The upside for NV is the industry hardly uses these features for now. The downside for AMD is that they bet too early on these features and the industry is far behind in their adoption.

4) AMD publicly stated that they have no interest whatsoever to pursue low volume design wins. With Titan's launch rumored at just 10,000 units, AMD has better things to do than waste hundreds of millions designing a 500-550mm² die GPU that less than 0.1% of GPU consumers will buy. It doesn't make any financial sense. For NV it's about selling K20X scrap rather than throwing it out completely. The Titan for NV is also about building brand equity and reinforcing the idea that they are a class leader in graphics. After-all, AMD is primarily a CPU/APU/server company (in Q4 alone AMD's CPU computing division lost more than than their entire graphics division made in 2-3 years). Contrast this with NV which is primarily focused on visuals and graphics, only recently having expanded into other markets with Tegra, which now comprises 30% of their revenue. They have done so because the GPU market is hardly growing. 

http://www.brightsideofnews.com/news/2013/2/13/nvidia-geforce-titan-launches-february-18th2c-2013-loses-to-gtx-6902c-amd-hd-7990.aspx

5) Even when AMD launched HD5850-5870 and completely dominated NV for 6 months, it hardly made an impact on their profitability. Then when HD7970 launched and had class leading performance for 2.5 months, it hardly mattered. This is more evidence if anything that these high-end GPUs do not really matter. Most of the GPU sales happen in the $100-250 price segment, not in the $400-500 ones and definitely not in the $800+ levels. 

Nvidia's own GPU breakdown shows that very few people buy GPUs above $199 level. The amount of total GPUs that are sold in the $400+ level is less than 5% of NV's/AMD's total discrete GPU sales.

6) The high-performance OEM market for companies like HP/ Dell / Lenovo also shows the picture clearly. Less than 7% of all desktop PCs sold are high-end systems, using at minimum a Core i5 or FX8000 series CPUs. The definition for this price segment is "Premium prices in the range of $1000 and up are usual in this category, and towers and minitowers are the most common form-factors. Processors used here include Intel’s Core i7, Core i5 or AMD’s FX and A10 chips."

http://www.xbitlabs.com/news/other/display/20121108145442_High_End_Personal_Computers_Account_for_Less_Than_10_of_PC_Market_Report.html

7) Rumors keep hinting that the design wins for PS4/720 are not just licensing of AMD's graphics IP but apparently AMD may have secured the delivery of actual chips that will go inside those consoles. The result is this new product stream (IPs and Other) may account for 20-40% of their total revenue by Q4 2013. It's a lot more strategic and financially lucrative to focus on delivering on these contracts to Sony and MS (not to mention the risks and obligations), rather than spending millions of dollars in a very highly competitive/contested $500+ discrete GPU market. Since NV lost the design wins for all 3 next gen consoles, they have to make up their profits somewhere, which ultimate meant raising the price of their products (GK104 is really a mid-range GTX560Ti successor but NV priced it at $499). 

So why would AMD waste millions of dollars pulling a similar PR stunt as the Titan when the Titan:

1) Will hardly matter for NV in terms of making $, but it makes sense for them instead of throwing out K20X scrap;

2) Changes nothing about AMD still retaining class-leading performance and price/performance on the desktop at nearly every level from $100 to $500;

3) Misses the point that AMD lost the most market share in the mobile space, not desktop space, which suggests AMD's greatest area where NV beat them was GTX600M parts (NV reportedly secured 300+ design wins with Kepler while AMD dropped the ball refusing to do custom design wins and go aggressively after the mobile OEM markets).

No matter how you slice it, the Titan is an amazing enthusiast card, but overall, it means squat for the graphics market. 

P.S. And finally even if you are a serious enthusiast (i.e., you don't mind tinkering), you will think twice or three times before wasting $900 on a Titan because 3x HD7970s make $250 a month bitcoin mining (!). For those who research bitcoin mining, you can make $ on 3x $350 HD7970s in the foreseable future. The Titan still seems like the most financially irresponsible product to purchase since it will be slower and will depreciate over the next 12 months, making $0 in the process towards next generation 20nm upgrade path. When HD7970s are fully paid off and will have made $1000-1,500 towards the next HD8970 x 3 upgrade, in that context the Titan doesn't look so hot.

See my reply below. Regarding NV's move with GTX700 series, those cards have already been rumored to launch only by Q4 2013. This was reported weeks ago, way before the Twitter feed on HD8000 series. Sources were already hinting at no new 28nm parts (other than the Titan) from either AMD or NV until Q4 2013. 

http://www.tomshardware.com/news/Radeon-GeForce-Delay-GPU-Next-Generation,20838.html

Due to selective journalism, the rumored timeframe for GTX700 launch hasn't been discuss despite the same sources pointing out on the same day that both HD8000 and GTX700 cards are likely MIA until Q4 2013.

No such confirmation was issued. This is a spin by news media for hits / regurgitation of old rumors that incorrectly predicted that desktop HD8000 parts would launch in the first half of 2013. AMD themselves never officially said when HD8000 desktop parts would launch in the year 2013. They only said they would launch in 2013, nothing more.

CES road-map for AMD last year:

Updated road-map this month shows HD7000 will be stable until Q3 2013, but it doesn't preclude AMD from launching HD8000 desktop parts in Q4 2013:

http://videocardz.com/images/2013/02/AMD-Radeon-January-2013-Slide-2.jpg

"I should note that [desktop] HD 8000 Series has never been so much as hinted at for a channel release. Anything to the contrary is an unsubstantiated rumor fabricated to drive traffic.” – AMD's Robert Hallock

More news came out this week that contradicts the rumors to begin with but the same media sites that published the original articles for hits would never admit they screwed up and fabricated/falsified info based on lack of any hard evidence.

PCWorld reached out to AMD in the wake of all the hub-bub to get a feel for the situation, and while there are still plenty of questions up in the air, I was told by a chortling AMD representative, "We will certainly have new [GPU] products in 2013." The company plans to clarify its 2013 plans for the Radeon brand later this week.

http://www.pcworld.com/article/2027813/amd-promises-to-squash-confusion-and-quickly-clarify-2013-radeon-plans.html

Doubtful. All rumors thus far have pointed to GTX780 (GK114) and HD8970 being mild-refreshes with 15-25% more performance over their 28nm parts. The Titan appears to be the exception, serving as an Enthusiast's Limited Edition part. The monster cards you are talking about are probably 20nm Volcanic Islands and Maxwell parts. It's unthinkable that HD8970 would have 7 billion transistors on 28nm node when HD7970 Ghz uses more than 230W of power in games.

Official specs/card pictures will leak February 19th and official reviews are slated for February 21st.

Only 10,000 cards are rumored to be made for the entire world.

Here are some early leaks.

The card may have voltage Unlock/control through the advanced control panel

Pictures of the card itself

No backplate

Real world power consumption should come in below 250W.

6 power phases, up from 4 on the GTX680.

More advanced version of GPU Boost 2.0 where the fan speed is now controled not only by temperatures but dynamic voltages.

Adaptive Vsync works up to 80 fps now, instead of 60 fps