The 1.1V voltage is part of the reason for this improved efficiency, as noted by it being better than the typical 1.2V. A further focus on reducing heat with using new package materials is another factor, along with an optimized circuit design. This leads to a staggering 70% decrease in thermal resistance when compared to the previous GDDR6. If you've owned something like an NVIDIA GeForce RTX 3090, you'll know that GDDR6 VRAM can get toasty under heavy usage.
VRAM has been an important selling point of modern GPUs, more-so than ever before. With the emerging use of technologies such as ray tracing, path tracing, and overall higher textures, VRAM hungry titles abound. Add to the proliferation of higher resolution monitors hitting more salient price points for consumers, and it is a recipe for high VRAM usage.
This will boil down to not only the amount of VRAM, but also its speed and efficiency, as
teased by Samsung. Technologies such as NVIDIA's DLSS 3 with frame generation further help to extrapolate performance from the latest GPUs, while also reducing power usage to some extent. This will likely be a key component in not only the ultra powerful segment of NVIDIA RTX 50 series GPUs, but the entry level models as well.
With GPU pricing of existing
GeForce RTX 40 Series products fluctuating over time and recently decreasing somewhat, pricing for GeForce RTX 50 series products will be watched with a keen eye by gamers. The inclusion of faster, more efficient memory will certainly make a big impact on how these new GPUs are received, in performance and in the pricing sensibilities of gamers.