Skip to main content

Nvidia’s new RTX 4070 model will use slower GDDR6 memory but won’t drop in price.

Markets of all kinds are being hit with a practice known as “shrinkflation,” where the price of an item remains the same but you get less product (or a lesser-quality product) for what you pay.

Shrinkflation is what comes to mind with the news that Nvidia’s latest version of the RTX 4070 graphics card will downgrade its 12GB of memory from GDDR6X to GDDR6.

That switch is a little less dramatic than it might sound. Despite the older and slower memory standard, the new RTX 4070 model will only lose about 5 percent of its memory bandwidth. That brings the memory speed down from 21Gbps to 20Gbps, according to VideoCardz.com.

And that’s just the memory. Every other component will remain the same, including the graphics processor and its clock speed.

So, the final effects on performance for the new model versus the old one will be negligible. GPU suppliers aren’t even updating their SKUs, though they are marking the new one with a “GDDR6” label. If you’re specifically hunting for GDDR6X memory, as will still be used on more expensive Nvidia cards, you might be in for a bit of a retail hunt.

The thing that really irks me, though, is that the price for the newer, slightly slower cards isn’t charging. Nvidia says it’s making the change “to improve supply and availability to meet strong demand,” according to two paragraphs buried in a press release about Black Myth Wukong and Star Wars Outlaws. Nvidia seems to be streamlining its supply chain for slightly better efficiency… but that’s a benefit for Nvidia and its suppliers, not for us consumers.

Shrinkflation, whaddayagonnado. If it’s any consolation, graphics card prices are so malleable by the time they get to retailers that we likely wouldn’t have noticed any change in the MSRP anyway.