AMD's Full Teaser Text

On June 01, 2016 at 10 a.m. China Standard Time (3 a.m. BST / 4 a.m. CEST) the Radeon Technologies Group will be announcing:

  • Radeon™ RX 480 set to drive premium VR experiences into the hands of millions of consumers; priced from just $199
  •  First Polaris architecture-based graphics processor to deliver VR capability common in $500 GPUs; expected to accelerate the size of the VR-ready install-base and dramatically increase the pace of VR ecosystem growth
  • RadeonTM RX 480 specifications including:
  AMD Radeon RX 480
TFLOPs (FMA) >5 TFLOPs
Compute Units 36
Memory Bandwidth 256GB/sec
Memory Clock 8Gbps GDDR5
Memory Bus Width 256-bit
VRAM 4GB/8GB
Typical Board Power 150W
VR Premium Yes
AMD FreeSync Yes
DisplayPort 1.3/1.4 HDR

Set to formally launch on June 29th, the Radeon™ RX 480 will deliver the world’s most affordable solution for premium PC VR experiences, including a model that is both HTC™ Vive Ready and Oculus™ Rift™ certified and delivering VR capability common in $500 GPUs.

In a notable market survey, price was a leading barrier to adoption of VR. The $199 SEP for select Radeon™ RX Series GPUs is an integral part of AMD’s strategy to dramatically accelerate VR adoption and unleash the VR software ecosystem. AMD expects that its aggressive pricing will jumpstart the growth of the addressable market for PC VR and accelerate the rate at which VR headsets drop in price:

  • More affordable VR-ready desktops and notebooks: AMD expects that affordable PC VR enabled by Polaris architecture-based graphics cards will drive a wide range of VR-ready desktops and notebooks, providing a catalyst for the expansion of the addressable market to an estimated 100 million consumers over the next 10 years.
  • Making VR accessible to consumers in retail: Thus far, retail has not been a viable channel for VR sales as average system costs exceeding $999 have precluded VR-ready PCs from seeing substantial shelf space. The Radeon™ RX Series graphics cards will enable OEMs to build ideally priced VR-ready desktops and notebooks well suited for the retail PC market.
  • Unleashing VR developers on a larger audience: Adoption of PC VR technologies by mainstream consumers is expected to spur further developer interest across the ecosystem, unleashing new VR applications in education, entertainment, and productivity as developers seek to capitalize on the growing popularity of the medium.
  • Reducing the cost of entry to VR: AMD expects that affordable PC VR enabled by Polaris architecture-based graphics cards will dramatically accelerate the pace of the VR ecosystem, driving greater consumer adoption, further developer interest, and increased production of HMDs, ultimately resulting in a lower cost of entry as prices throughout the VR ecosystem decrease over time.

The Radeon™ RX Series launch represents the first salvo in AMD’s new “Water Drop” strategy aimed at releasing new graphics architectures in high volume segments first to support continued market share growth for Radeon™ GPUs. In May 2016, Mercury Research reported that AMD gained 3.2% market share in discrete GPUs in Q1 2016. The Radeon™ RX Series will address a substantial opportunity in PC gaming: more than 13.8 million PC gamers who spend $100-300 to upgrade their graphics cards, and 84% of competitive and AAA PC gamers. With Polaris architecture-based Radeon™ RX Series graphics cards, AMD intends to redefine the gaming experience in its class, introducing dramatically improved performance and efficiency, support for compelling VR experiences, and incredible features never before possible at these prices.

Supporting Quotes:

“VR is the most eagerly anticipated development in immersive computing ever, and is the realization of AMD’s Cinema 2.0 vision that predicted the convergence of cinematic visuals and interactivity back in 2008,” said Raja Koduri, senior vice president and chief architect, Radeon Technologies Group, AMD. “As we look to fully connect and immerse humanity through VR, cost remains the daylight between VR being the purview of the wealthy, and universal access for everyone. The Radeon™ RX Series is the disruptive technology that adds rocket fuel to the VR inflection point, turning it into a technology with transformational relevance to consumers.”

“The Radeon™ RX series efficiency is driven by major architectural improvements and the industry’s first 14nm FinFET process technology for discrete GPUs, and could mark an important inflection point in the growth of virtual reality,” said Patrick Moorhead, principal analyst, Moor Insights & Strategy. “By lowering the cost of ownership and increasing the VR TAM, Radeon RX Series has the potential to propel VR-ready systems into retail in higher volumes, drive new levels of VR content investment, and even drive down the cost of VR headsets.”

“We congratulate AMD for bringing a premium VR ready GPU to market at a $199 price point,” said Dan O’Brien, vice president of virtual reality, HTC.  “This shows how partners like AMD survey the entire VR ecosystem to bring an innovative Radeon RX Series product to power high end VR systems like the HTC Vive, to the broadest range of consumers.”

AMD Teases Radeon RX 480
Comments Locked

377 Comments

View All Comments

  • Ranger1065 - Wednesday, June 1, 2016 - link

    Oh dear. With no sign of Chizoo, I hoped Nvidia trolls were on the decrease @ Anandtech.
  • Zingam - Wednesday, June 1, 2016 - link

    I agree... I prefer cheaper! :)
  • gilmour - Wednesday, June 1, 2016 - link

    Pretty sure some variation of this card will be my next purchase due to price and my performance needs but still disappointed in the specs that were shown.
    Considering the massive difference in clock speeds the GTX1070 can do at 150W it's very surprising how low the RX480 is at 150W.
    Either the specs shown are still concealing real clock speeds and power draw and are better than shown, or the RX480 is a low end Polaris 10? part and there is a lot of headroom to release faster 480 and higher specc'ed 490 parts albeit with increased power draw.
    Still seems Nvidia did some impressive engineering to get the clock speeds, performance and power draw at the levels they did, AMD maybe not quite so good.
  • beginner99 - Wednesday, June 1, 2016 - link

    You can't compare TDP from NV and AMD. NV usually greatly understates TDP as was the case with 970 and 980. In real games the power use difference of whole system between a 970 and 290 was much less than the difference in TDP.
  • Yojimbo - Wednesday, June 1, 2016 - link

    TDP is not a measure of average power consumption. The TDP rating indicates the maximum sustained heat the chip is allowed to dissipate in real world usage, or put another way, the minimum heat the cooling system must be able to dissipate. If NVIDIA were to "greatly understate TDP" then their cards would be burning up left and right because their cooling systems wouldn't be able to keep up.

    If a card were to be consistently closely approaching its TDP across a wide range of uses then I think that card would probably be TDP bound. If one card is far from its TDP in a particular sustained real world use case, then there must be some other real world use case where it uses a lot more power, because otherwise the TDP claimed for the card could be lowered.

    As far as what you said about the GTX 970 in comparison to the R9 290, it seems to me that you probably saw figures in just a small number of use cases. But if the 970 really is able to more consistently approach its TDP over an average of a large range of uses it seems to me that it implies the underlying architecture of the GTX 970 is more balanced, because there must be more extreme outliers for the R9 290 to require its TDP to be upped for those cases.

    Regardless, the R9 290 consistently used significantly more power than the GTX 970.
  • Yojimbo - Wednesday, June 1, 2016 - link

    An interesting thing I've noticed is how much NVIDIA's transistors per core count has gone up from Kepler to Maxwell and also from Maxwell to Pascal while AMD's transistors per core count has been relatively flat from 1st through 3rd generation GCN. I wonder what the transistors per core situation is for 4th gen. GCN.
  • ajlueke - Wednesday, June 1, 2016 - link

    Well, in Maxwell to Pascal you have actually changed process nodes, so you expect the number of transistors that can be placed on a die of the same size to increase. Now from Kepler (GTX 680) we went from 3.5 billion transistors to 8 billion in Maxwell 2 a 128% increase, all on the 28 nm process. While, AMD went from the 4.3 billion transistors in the HD 7970 (GCN 1) to 8.9 billion in the R9 Fury X (GCN) and increase of 107% all on the 28 nm process. Doesn't seem that different to me.
  • Yojimbo - Wednesday, June 1, 2016 - link

    I'm not talking about transistors per die area, but transistors per compute core (CUDA core as NVIDIA calls it). What that means is that NVIDIA has been using a greater percentage of the total transistors for cache, register files, ROPs, schedulers, and the like, and a lesser percentage on ALUs. But they have been getting greater efficiency out of their architectures by doing so.
  • pashhtk27 - Wednesday, June 1, 2016 - link

    Finally a card that I can dream of buying, something that won't make the poor me feel too bad investing on.
    If it can really offer gtx970 level performance, I'm in. Looking forward to release and benchmarks.

    Will make a nice combo with egp docks when cheaper ones come out.
  • webdoctors - Wednesday, June 1, 2016 - link

    I have the same thought/question. How does this card compare to the 970 which is now going for about the same price? Do we have a new $199 king?

Log in

Don't have an account? Sign up now