Just a few days before next-gen AMD (AMD - Get Report) gaming GPUs go on sale, Nvidia (NVDA - Get Report) is unveiling a pretty strong counterattack that could help it maintain its lead in the high end of the market.
On Tuesday morning, Nvidia confirmed recent media reports by announcing the GeForce RTX 2080 Super, 2070 Super and 2060 Super, a trio of new gaming GPUs that feature suggested graphics card prices of $699, $499 and $399, respectively. The 2060 Super and 2070 Super will both go on sale on July 9, while the 2080 Super will do so on July 23.
As their names suggest, the "Super" GPUs are more powerful versions of Nvidia's standard GeForce RTX 2080, 2070 and 2060 GPUs, thanks in part to the packing of additional GPU cores. Nvidia claims that on average, the 2070 Super and 2060 Super will respectively be 16% and 15% faster than their predecessors.
As of mid-day, Nvidia's shares are down 1.5% in Tuesday trading to $163.75, amid broader declines for chip stocks. AMD's shares are down 0.8% to $30.95.
In addition to having more cores, the 2080 Super features slightly more memory bandwidth than the standard RTX 2080, and the 2060 Super will come with 33% more memory bandwidth and 2GB of additional graphics memory (8GB vs. 6GB) relative to the standard 2060.
Nvidia's new GPUs at a glance. Source: Nvidia.
With the 2080 Super and 2070 Super featuring the same suggested prices as what the standard 2080 and 2070 have featured to date, Nvidia will stop selling the standard versions of the GPUs. However, the company will still offer the standard 2060, which at a suggested price of $349 is $50 cheaper than the 2060 Super. And at the very high end of its gaming GPU lineup, Nvidia will still sell its RTX 2080 Ti GPU, which has a suggested price of $999.
Reviews of the Super 2070 and Super 2060 that were published on Tuesday morning feature test results that differ slightly from Nvidia's performance claims, but not massively so. In an 11-game test, PC Gamer found that on average, the Super 2070 and Super 2060 respectively outperformed the standard 2070 and 2060 by 9% and 15% when playing games at a 1440p resolution and Ultra quality settings. At a 4K resolution, average gains were respectively 10% and 21%.
AnandTech's review of the GPUs, which relied on a test system with a faster Intel (INTC - Get Report) CPU, turned up somewhat better performance gains for the 2070 Super, with reviewer Ryan Smith concluding that both the 2070 Super and 2060 Super are "about 15% faster than their direct, non-Super predecessors." He notes the 2060 Super's performance is similar to that of the standard 2070, which had been selling for $100 more, and that the 2070 Super nearly matches the performance of the standard 2080, which had cost $200 more.
During a press briefing, an Nvidia spokesman admitted the performance gains delivered by the Super GPUs relative to their predecessors will be accompanied by roughly proportionate increases in power consumption. However, he also stressed that the Turing GPU architecture that both the Super and standard RTX products rely on compares favorably to AMD's new Navi GPU architecture in terms of power efficiency.
AMD is three weeks removed from unveiling its first Navi GPUs, the Radeon RX 5700 XT and 5700. Graphics cards featuring the chips go on sale on July 7th, and will respectively be priced at $449 and $379. As a result, the GPUs were positioned as rivals to the standard RTX 2070 and 2060. At the time that they were revealed, AMD showed benchmarks that had the 5700 XT, which has a graphics card thermal envelope (TDP) of 225 watts, slightly outperforming the standard 2070, which has a TDP of 175 watts, when running games at a 1440p resolution. Likewise, tests showed the 5700, which has a TDP of 180 watts, slightly outperforming the standard 2060, which has a TDP of 160 watts.
Judging by what reviews show about their performance and power draw relative to the standard RTX 2070 and 2060, Nvidia's Super 2070 and 2060 should respectively be able to moderately outperform AMD's 5700 XT and 5700 when performing traditional graphics rendering, while remaining competitive in terms of power consumption. And like other Turing GPUs that have the RTX label, the chips will also be differentiated through their inclusion of specialized processing cores for handling real-time ray tracing (a demanding rendering technique that can enable photorealistic game imagery) as well as a technology known as deep learning super sampling (DLSS), which involves using AI/deep learning algorithms to boost a game's frame rates.
Current and upcoming games supporting real-time ray tracing. Source: Nvidia.
Between the performance gains delivered by its Super GPUs and its ability to use real-time ray tracing and DLSS as selling points, Nvidia looks poised to maintain its current, dominant position in the high-end and upper-mid-range segments of the gaming GPU market -- particularly as more games supporting ray-tracing and DLSS arrive on the scene.
This article has been updated to mention reviews of Nvidia's Super 2070 and Super 2060 GPUs, and to note Nvidia and AMD's share price movements on Tuesday.
Save 57% with our July 4th Sale. Join Jim Cramer's Action Alerts PLUS investment club to become a smarter investor! Click here to sign up!