After an earlier vague deadline of Q2’15 and more than a few teases in the interim, AMD has finally announced when they’ll be revealing their forthcoming high-end video card.

AMD will be hosting an event on June 16th at 9am PST to release the details on the card, in a presentation dubbed “AMD Presents: The New Era of PC Gaming.” The presentation will be taking place at the Belasco Theater in Los Angeles, CA during E3 week, and happens to be where the AMD-sponsored PC Gaming Show also takes place that evening. This event is open to the public, or can be viewed via webcast.

I would quickly note here that at no point does AMD specifically call this a launch. And for that matter, the last time they held a public event like this – the Radeon 200/Hawaii reveal – Hawaii didn’t launch until a month later. In this case AMD has already committed to a June launch for the card, but at the moment we’re not expecting to see the card go on sale on the 16th.

Update: As part of AMD's Computex event (was was almost entirely focused on APUs), AMD CEO Dr. Lisu Su also briefly teased the 16th event, pulling out AMD's new high-end GPU. The chip being shown is a complete package with its silicon interposer and organic substrate. The crew over at Tom's Hardware got a good picture of it.

Image Courtesy Tom's Hardware

Source: AMD



View All Comments

  • haukionkannel - Wednesday, June 3, 2015 - link

    You have to remember that AMD R&D budget is very tight at this moment. HBM memory is new and expensive at this moment, so that is a good reason to make only high margin niche high end product this time. So that they can test that HBM memory in practice.
    We also know that 2016 we will at last get 1xnm products from Nvidia and AMD. If AMD would use its development budget this year to make new 28nm GPU, there would not be money next year for 14nm GPU development... So AMD has to make hard choices. New card now with 28nm, or new card next year with 14nm production node. When you consider that, it is no brainier decision for AMD to make re brands this year and leave new cards to next year.
    So why Nvidia make new GPUs then? Well because they can. The have much bigger development budget than AMD has at this moment, so they can make new chips every year, much like Intel can do in CPU part. If AMD would have had money, we would have got to Zen a long time ago in CPU part. Also we would have seen full HBM upgrade tour even in this year. But when the money is tight, you have to make good decisions and leaving more money to 14nm products is the smart move at this moment!
  • nunya112 - Wednesday, June 3, 2015 - link

    well we already know that around CES. AMD revised the silicon. that is fact! as Titan was released, and AMD announced the first delay essentially. Went radio silent.
    and then Lisa Su announced that Computex was the release for the new chip. We know this didnt happen. and a few things really got under my skin here.

    The EA/DICE employee the posted that picture. was obviously a media ploy by AMD. AMD confirmed at computex that they dont even have bioses or clocks etc nailed down. and AMD didnt have a full unit for display even!! So that was a fake!! I'm very upset at this.

    2nd. They missed another deadline here. and at E3 they will miss another. you cant have a card that far behind be ready in what 13 days! That's impossible! Remember they said they would be FOR SALE! There is NO WAY! Again another deadline. and really under my skin.

    leaks coming out of Computex are. the AIB's were given a demo of an unfinished Fiji/ FURY whatever. and it performed worse than a Ti 980. So I would have hated to see the A revision. cause not beating a 980 vanilla. that wld have been really bad.

    Instead realistically at the most we will see Fiji run as well as a 295 X2. and thats if they get every single thing 100% right! drivers, bios, water cooling. everything.

    So realistically we may even see 85% of a 295X2 speed.

    For perspective a Titan X is about 20-30% slower then 2 780Ti's in SLI.
    And we all know they all use the same node at TSMC. or maybe AMD have global foundries making it. who knows then that makes things a bit harder to judge!
  • Kinglock - Wednesday, June 3, 2015 - link

    Wait a minute! Are they going to launch only the high end gpu that will cost around $600?? Cause if thats the case i go for a gtx 970 since my budget doesnt go that high!! Reply
  • Kinglock - Wednesday, June 3, 2015 - link

    And another thing, if they relalease mid range gpus but having them rebranded i still wont buy them since AMD previous architecture is crap(power hungry and gets hot) Reply
  • milkod2001 - Wednesday, June 3, 2015 - link

    it's less then 2 weeks to find out :) Whatever comes from AMD,there's good chance existing prices will drop slightly so you will be able to get 970 for less.

    i do not expect any breathtaking changes in mid range GPUs until 14/16nm maybe next year
  • ES_Revenge - Saturday, June 6, 2015 - link

    This is and always will be the problem with "duking it out in the high end" you get bragging rights and cool halo products but I can't imagine you're raking in money hand over fist in sales, unless these things also have huge profit margins (which they don't given all the R&D).

    Because iGPUs/APUs have removed a lot of the low-end from the graphics card market, that leaves the main market being the mid-range cards. From $150-250. What I never understood was why the cards in this segment have been absolutely stagnant in terms of pricing and have only gotten worse over time.

    You had the HD 7950 and 7970 which became the upper mid-range when Hawaii dropped as the 290 and 290X. At that time you could get 7950s at ~$200 and 7970s at ~$270 (and really for most the additional money for the XT was still not worth it). But the re-brands, the 280 and 280X sold for much more and continue to sell for more today...even now, years later.

    So today you have 280Xs, 285s, and GTX 960s going for $250 US (and like $300 Canadian), for performance that was high-end 3.5 years ago, and at upper mid-range for the past year and a half or so. So while all this nonsense with GTX 970, 980, 960, and 980 Ti (which really should have been called Titan Jr. or something) has been going on, the mainstream has had to sit by and watch prices *increase* and performance go absolutely nowhere.

    Seems we pay for the high-end, even if we're not paying for it directly. And most people on these sites seem to care mainly about these $600-$1500 cards even if they're not buying them, themselves. Obviously you can't have lower end cards based on the same tech without the high-end first--the high end always launches first---BUT this whole time the mid-range has been sitting doing mostly nothing *for a while* and it's pretty stupid IMO. After all, this is where the money is made for these companies, right?

    Even with AMD being a solid choice in the $150-250 range, they still seem to be doing poorly financially and still have the need to chase the high end without even considering changing anything in the mid-range. I'm all for it if the new Fiji cards are $600+, *if* we also get Hawaii-level performance at the $200 mark along with it. However that seems to not be the case here. Instead we just keep on trucking with the same performance and prices higher than before, for no reason at all, other than to finance high-end projects for the minority that buy these cards.

    Nvidia hasn't done anything great either. GTX 960 is swell because of it's relatively low-power envelope, HEVC features, etc. but at the price it sells for? For the same performance we had for less money, nearly 2 years ago? As nice as the GTX 970 is for the price, it's still at $350 or so in absolute terms, when IMO we should have that kind of performance at the $200 mark by now. Similarly I'm all for the 980 Ti, if 970s are selling for a lot less than they are.

    Interestingly the used market doesn't really cater to these ridiculous prices. While Nvidia's 900-series seem to do okay (compared to retail) because of their relative "new-ness", AMD's R9 cards do not. Try getting more than $150 for a 280, or more than $200 for a 280X. Not likely, even if it's brand new won't get you anywhere. 270/270Xs frequently sell in the $100-120 range as well, which is probably where the best bargain is right now (two of these used for $200, in Xfire, are about as fast as a 290, albeit possible FTV issues).
  • D. Lister - Monday, June 8, 2015 - link

    I understand your complain regarding the prices these days. Especially in instances like Nvidia launching, which for all intents and purposes is, a Ti gpu as a $1,000 Titan X.

    <rant> Sorry Nvidia, it is an awesome gaming GPU, but no DP = no Titan, period. What? You say it's because Maxwell arch. doesn't have DP capability? Aww... well, that's YOUR problem. By selling a Ti (with useless ammount of RAM tacked on) as a Titan, you have actually damaged the Titan brand. Remember, the original Titan gpu was named after the Titan Supercomputer, which used the NV GPU for its DP ability. So sell a Ti as Titan and then call a downscaled chip a Ti? That's what we call a "dick move", if anything. </rant>

    Anyway, the abovementioned shenanigans aside, performance per dollar has improved immensely. A modern $300 GPU would destroy a $300 GPU from 10-15 years ago, regardless of their tiers.

    TLDR, the occasional spike notwithstanding (295X2, TZ, TX, etc.), we get a much bigger bang for our buck today, than we did a decade ago. So all is well... I guess. :)
  • wow&wow - Wednesday, June 3, 2015 - link

    "... the last time they held a public event like this – the Radeon 200/Hawaii reveal – Hawaii didn’t launch until a month later."

    Then the CEO is a Ken, but now the CEO isn't a Barbie.
  • Shinku_sai - Friday, June 5, 2015 - link

    Evidently "Barbie" has a bigger package than "Ken" as well.... 😀 Reply
  • ruano23 - Wednesday, June 3, 2015 - link

    The announcement it's not about really about a new GPU, but a new interface implementing HBM memory. The rest is probably a little refinement of the GCN design.
    The main thing is HBM memory. If the new GPU used the GDDR5, it wouldn't be any big step.

Log in

Don't have an account? Sign up now