Advanced Micro Devices (AMD) has been on a roll as of late as it has managed to successfully push out its Polaris platform on the GPU (Graphics Processing Unit) market as well as its Zen CPU (Central Processing Unit) architecture as part of its new Ryzen platform and has begun to inch back market share in both the CPU market dominated by Intel and the mainstream GPU market even as its RX 400 and RX 500 (refreshed RX 400) series continue to boast impressive demand backed, at least partially by Cryptocurrency miners and enthusiasts who have found, over time that Polaris-based GPUs have the lowest cost of entry as well as a low performance per watt rating in the GPU mining business. AMD’s stock has rallied accordingly; it is up nearly 140% over the past year and investors remain bullish about the stock that used to trade as low as $2 in Jan 2016 plagued by falling market share and limited avenues to generate revenue in the short term.
There is however a lucrative market that AMD has not been able to crack, one that continues to fuel much of rival Nvidia Corporation’s profits: Enthusiast Gamers. Nvidia’s foray into the consumer graphics market was a high-profile entry of the GTX 1070 and GTX 1080, GPUs that were sold at a premium from their AIB alternatives as part of Nvidia’s “Founders Editions” program which saw Nvidia price gouge the fact that its GPUs were not available with AIB partners in any other form in the weeks that followed. A year later, Nvidia churns out a new king of the hill, the GTX 1080TI which is slightly more powerful than the incumbent Titan X Pascal, the highest end consumer-grade GPU (This would later be replaced with the now incumbent Titan XP, the fastest consumer-grade GPU in the market). While we did get an impressive launch from AMD vis-à-vis its RX 400 series in the same year which aimed to take control of the mainstream market, a task it has performed admirably, we have seen nothing from the California-based silicon maker apart from the whispers of the codename for its new GPU architecture, the Radeon RX Vega.
AMD’s delay in launching Vega has served the enthusiast market to Nvidia on a silver platter. However, we do see light at the end of the tunnel for AMD as it reaffirms its rollout of Vega in the latter half of June this year as confirmed by AMD CEO Lisa Su. We also that the first GPU to launch from AMD in the Vega line-up will by the RX Vega Frontier Edition (similar to Nvidia’s own Founders Edition branding) and will contain 16GB of HBM2 memory and come with approximately 13 TFLOPs of Single Precision raw compute power (FP32). To put matters into perspective, Nvidia’s current top-of-the-line GPU, the GTX 1080TI has 11.3 TFLOPs of FP32 compute power. While this is not the only play when it comes to GPU’s (optimization plays a large role), this does prove that AMD is pulling no punches when it comes to pushing out Vega with a flagship product in tow. The gaming edition which will launch later this year is expected to even eclipse these projections in terms of raw power. According to Raja Koduri, Chief Architect Radeon Technologies Group at AMD, it will be cheaper and faster for enthusiast gamers. With 16GB of HBM2, AMD’s Frontier Edition Vega will at the very least be priced at a far higher rate than most mainstream gamers can afford to play. Given speculation that there is a supply-side limitation due to AMD’s use of the high-price HBM2 memory that has supply restrictions in play, it makes sense that a company like AMD would stagger the launch of Vega, opting to push out the highest end GPU first and follow it up with lower cost variants as time goes on. While AMD does not have a hard launch date in play for many territories, we expect to learn a lot more about its RX Vega hardware in the coming days as well as see the RX Vega in action towards the end of June or early July as it rolls out to reviewers for benchmarks courtesy of AMD.