Will A GDDR6X Shortage Lead To A GeForce RTX 4070 With Downgraded Memory? Not So Fast

NVIDIA's most powerful GeForce RTX graphics cards use GDDR6X memory as offers more stable signaling, in turn granting higher transfer rates—which is exactly why NVIDIA uses it. However, because the company's graphics cards are the only things that use GDDR6X memory—and the next-generation cards will use lightning-fast GDDR7 RAM—production of GDDR6X was limited, and it appears to have stopped.


This absolutely isn't unthinkable; NVIDIA does like to launch odd variants of extant GPUs late in the product life cycle. We've seen this before with cards like the 8GB GeForce RTX 3060 and the late launch of the Ampere-based GeForce RTX 2050. The site speculates that the card would be based on an even-further cut down version of the AD104 GPU that powers every other "RTX 4070" card besides the GeForce RTX 4070 Ti Super.

We're not completely confident in the idea that NVIDIA will launch another RTX 40-series card before the RTX 50 series, though. By all accounts, there's already ample stock in the channel; adding another SKU would only overload sellers that are already struggling to move expensive graphics cards. Furthermore, there's the fact that this is simply speculation on Benchlife's part based on a rumor that production of GDDR6X memory has stopped. It makes sense, but there's no proof just yet.

According to prior rumors, NVIDIA's top-end Blackwell-based GeForce GPU will be targeting a 48% performance uplift from the GeForce RTX 4090. If that's true, it's absolutely going to need the big memory bandwidth gains to be had with GDDR7 RAM. Those parts are expected to show up around CES, if not earlier, so we don't have that long to wait.