In another quick shift in the hyper-competitive performance video card market, AMD sends word this afternoon that they are enacting some price cuts that will be taking effect later this week. This latest round of price cuts comes hot on the heels of last week’s launch of the GeForce GTX 660 Ti, which saw NVIDIA introduce their first 28nm performance video card at $299.

The bulk of the cuts here will be for the 7800 series, where the 7870 in particular is finding itself somewhat displaced after the launch of the GTX 660 Ti. The $299 660 Ti isn’t necessarily in direct competition with the already-cheaper 7870 – which had a street price of around $279 last week – and since AMD had already quietly shuffled prices around ahead of the GTX 660 Ti launch, we weren’t expecting any further changes here. But it would appear that the gap between the 7870 and GTX 660 Ti is closer than AMD would like.

As a result the 7870 will be getting a slight price cut to push prices towards $249. This would make the card a full $50 cheaper than the GTX 660 Ti, which is apparently the kind of leverage AMD wants right now.

Meanwhile because the 7870 is getting a price cut, so is the 7850. AMD is expecting the street prices on the 2GB 7850 to fall to around $209 after the price cuts take effect, putting it $40 below the newly repriced 7870. The 2GB 7850 has been averaging $239 in the past week, so this would represent a price cut of around $30. Meanwhile the extremely rare 1GB version of the card would end up below $200, though given how few of those cards exist it’s hard to say if it will hit AMD’s $189 price target.

Alongside those price cuts the 7800 series will be receiving a new game bundle promotion in a few weeks. The AMD Gaming Evolved title Sleeping Dogs will be AMD’s latest bundle, replacing the outgoing DiRT Showdown bundle. This will sit opposite NVIDIA's existing Borderlands 2 promotion, which went live last week. As with past bundles this is being done at a retailer level, so it’s primarily geared towards online retailers (e.g. Newegg) that can quickly bundle vouchers with new cards.

Second Summer 2012 Radeon HD 7000 Series Price Cuts
Card Launch Price Spring MSRP Summer MSRP Second Summer MSRP
Radeon HD 7970GE $499 N/A N/A $499
Radeon HD 7970 $549 $479 $429 $429
Radeon HD 7950 $449 $399 $349 $319
Radeon HD 7870 $349 $349 $299 $249
Radeon HD 7850 $249 $249 $239 $209
Radeon HD 7770 $159 $139 ~$119 ~$119
Radeon HD 7750 $109 $109 ~$99 ~$99

Meanwhile, along with the 7800 series the 7950 is also technically getting a price cut. We say “technically” because AMD seems to be rubber stamping price cuts that have already happened. The 7950 has been readily available below its $349 MSRP for quite some time now, and AMD’s new MSRP of $319 reflects the price of cards that are already available.

Finally, it should be noted that despite AMD’s official announcement we wouldn’t be all that surprised if only a few cards ended up reaching these new MSRPs. AMD lists their MSRPs as “starting at”, which means that AMD is listing the price of the cheapest card. This is largely how the previous round of price cuts played out, so pickings right at these new MSRPs may be slim.

Post-Cut Summer 2012 GPU Pricing Comparison
Radeon HD 7970 GHz Edition $469/$499 GeForce GTX 680
Radeon HD 7970 $429/$399 GeForce GTX 670
Radeon HD 7950 $319/$299 GeForce GTX 660 Ti
  $279 GeForce GTX 570
Radeon HD 7870 $249  
Radeon HD 7850 $209  




View All Comments

  • CeriseCogburn - Wednesday, August 29, 2012 - link

    The new core voltage on amd's failing process for the 79xx series means all the watt charts are too low.
    MOAR POWER, MORE $$$$$$$$, amd epic fail
  • Ananke - Wednesday, August 22, 2012 - link

    :) GTX660ti has asymmetric 196-bit bus, i.e. that thing effectively is a 128-bit card...GTX670 on the other hand is good hardware, but too expensive. Reply
  • CeriseCogburn - Thursday, August 23, 2012 - link

    Blown away verified owner doen't agree with your amd fanboy lies

    Pros: Out of the box very impressive. Card is well designed, lots of cooling added to the Twin Frozr. After some tweaking, I benchmarked 1330mhz stable while adding ~500mhz to memory as well. Max temp in BF3 overclocked everything Ultra was 76 degrees. Never dipped below 50fps from what I noticed.

    Overclocked, this card outperforms a 670 and is just barely shy of a 680 on benchmarks. $300 is worth every penny, plus with Borderlands 2 that technically drops it down to $240, making it more of that low-range card people were wanting out of it.
  • claysm - Tuesday, August 21, 2012 - link

    I use a 2560x1440 monitor... Reply
  • Flunk - Tuesday, August 21, 2012 - link

    High end graphics cards aren't really necessary unless you're driving screens that big. If they didn't put in those benchmarks everything would be CPU bound and therefore impossible to compare.

    I have a 1080p screen (23") driven by 460 GTX x2 and it's more than enough. You just don't need the GPU power at low resolutions.
  • RussianSensation - Tuesday, August 21, 2012 - link

    Good point. You can turn down a lot of settings in games such as Ultra in BF3 or Extreme in Crysis 2. That setting has a huge performance hit and barely any improvement in image quality. In fact in many games I can't even tell the difference between High and Very High/Ultra. Reply
  • CeriseCogburn - Wednesday, August 29, 2012 - link

    " RussianSensation on Tuesday, August 21, 2012
    Good point. You can turn down a lot of settings in games such as Ultra in BF3 or Extreme in Crysis 2. That setting has a huge performance hit and barely any improvement in image quality. In fact in many games I can't even tell the difference between High and Very High/Ultra. "

    Yet you go on about 8xmsaa on previous pages claiming amd wins....
    Which lie is it going to be next time ?
  • TheJian - Tuesday, August 21, 2012 - link

    None of the games in the 660TI review at anandtech are cpu limited at 1920x1200. The closest you come to that is Skyrim, which with 4xMSAA/16AF on, the spread of the cards is from 43.6 - 98.2fps. The difference between the Zotac AMP (92.8) and the Reference 670 GTX (98.2) is 5.4fps (~5%) and the 7950Boost only scored 86. Quite the spread. They'd all be pegged and within a fps of each other if CPU limited.

    Funny to note that the 660 is over 2x faster than the 560TI (43.6fps) it is replacing and that ryan says isn't worth the upgrade..LOL.
    From the conclusion page and part of what I attacked in my wall of text :) :

    "The GTX 660 Ti is actually a great upgrade for the GTX 560 Ti (and similar cards) from a performance standpoint, but despite the similar name it can’t match the GTX 560 Ti’s affordability. This entire generation has seen a smaller than normal performance increase at the standard price points, and the GTX 660 Ti doesn’t change this. If you’re frugal and on Fermi, you’re probably going to want to wait for whatever comes next."

    Nope not as cheap as the 560, but it's over two times faster in skyrim as shown. But you should wait for whatever comes next, because the 660TI just isn't a good improvement...LOL Smaller performance gain for entire generation? So a 112% performance improvement isn't worth the money? Small gain? When was the last time we replace the last model with a card that hits 100% faster? It's still going for ~$150. Double power, double price...Hmm...I don't get the conclusion I guess. It's 50-100% faster depending on game.
    560TI vs. 460 TI
    49.9 vs. 38.5 Civ5 29%
    61 vs. 44.8 Bad company 2 36%
    44 vs 33.1 Stalker CP 32%
    82.6 vs. 63.6 Dirt =29 %
    74 vs. 60.5 Mass Effect 2 =22%
    119 vs. 93 HAWKX =27%
    31 vs. 24 Metro 2033=29%
    There are more games, I just got bored.
    Pretty much 22-36% and I'm using 1920x1200 not the 1680x1050 scores which would make it less as the 460tI would be less stressed memory wise, so this is a best case for the gaps.
    There isn't a game in the list that doesn't improve by over 35% in the 660TI review vs. 560.
    Civ 5 660 TI review vs. 560 ti in same review
    51.9 vs. 35.9=44% So directly comparing to last gen 560 ti replacing 460ti, which only improved 29% over it in Civ 5 way back then, the UPGRADE to 660 is looking like a 15% improvement HIGHER than the previous gen...But Ryan thinks that 44% is LOWER than 29%. Where did he get his diploma? Skyrim shows 112% improvement...Again, 112% is worse than avg improvement of previous gen of 22-36%. I'm not sure what planet you are on if 112% is LOWER than 22-36%. Strange conclusions by ryan...ROFL.

    So monitors you should by from fly by night joints in Korea/New Zealand, 112% is less than 36% (or apparently 22% too...LOL), 44% is lower than 29% (really?) and the list goes on and on with the inaccuracies of this conclusion in the 660 TI article...There's just so much that is whacked. It was really to easy to pick apart. People may not like walls of text, but all should read page 4 of the All Comments section (in the 660 review) for his LAME excuses, and my rebuttal. It's really quite funny. I challenge him to debate it, ANY of it if he can. Heck, anyone take a shot at it...LOL. It's his benchmarks and his words. :) Incidentally RYAN did the review of the 550ti which I got the %'s from for the 560ti vs 460ti numbers...LOL. He can feel free to argue with himself too... :)
  • Blazorthon - Tuesday, August 21, 2012 - link

    Didn't affect a thing? That memory bus (beyond what people have said that proves you wrong about it anyway)and the associated hardware with it are the only difference between the 660 TI and a 670, if I remember correctly. The difference in performance between the two cards shows that the 660 TI's memory bus does in fact have a huge impact on its performance. The GTX 670 is already a memory-bandwidth bottle-necked card when its GPU is running code that it doesn't suck at (too much tessellation, DirectC, AA, et cetera) and the 660 TI proves it.

    If anything, AMD lowered prices because of not wanting to confront the hype/propaganda that is going around about the 660 TI, not because their cards are worse. They most certainly aren't anyway and I have no idea what is going on with Nvidia that would cause them to release such extremely unbalanced GPUs as the Kepler GPUs. Fermi, although hot, was a lot better as an architecture in many ways than Kepler and was also not kept so memory-bandwidth constrained. A die-shrink of it with minor reorganizing would have probably been better than Kepler.
  • RussianSensation - Tuesday, August 21, 2012 - link

    It's not just the memory bus width.

    660Ti has 24 ROPs vs. 32 ROPs for the 670. ROPs coupled with memory bandwidth is what helps to support MSAA performance.

Log in

Don't have an account? Sign up now