Nvidia’s plans for the refresh of the GeForce RTX 50 series of graphics cards has changed several times over the months, according to rumors. But yet fresh information seems to paint the RTX 50 Super GPUs as either outright cancelled, or at the least on hold and there’s 3 big reasons for all of this.
As a quick reminder, the leaked specs for the RTX 50 Super highlighted memory upgrades as the primary driver of an upgrade over the ‘vanilla’ cards,courtesy of 3GB memory modules. The memory bus of the various GPUs would remain identical, so for example the RTX 5070 would retain it’s 192-bit bus, but rather than 12GB on the ‘vanilla’ card, the 3GB GDDR7 modules would provide a total of 18GB VRAM.
Another key SKU was the RTX 5080, with the Super version of the card 24GB of GDDR7, with the same 256-bit bus the 16GB RTX 5080 offers. Combine that with potentially modest improvements in clock speeds and in some cases, a smattering of extra SM and you’ve an idea of what Nvidia were cooking up.
I’ll include the table for the previously rumored specs (most of which came from Kopite7Kimi on Twitter).
| SKU Name | SM Count | Memory Bus | RAM Config |
|---|---|---|---|
| RTX 5070 | 48 | 192 Bit | 12GB GDDR7 |
| RTX 5070 Super | 50 | 192 Bit | 18GB GDDR7 |
| RTX 5070 Ti | 70 | 256 Bit | 16GB GDDR7 |
| RTX 5070 Ti Super | 70 | 256 Bit | 24GB GDDR7 |
| RTX 5080 | 84 | 256 Bit | 16GB GDDR7 |
| RTX 5080 Super | 84 | 256 Bit | 24GB GDDR7 |
But, as I myself have reported several times according to my own sources – including in a recent video where I discuss the RTX 60 Specs and release date (with RTX 60 apparently launching 2H 2027), Nvidia have changed its plans.
According to Board Channels (though I first spotted this on GazLog), this is due to three factors. The first is memory – VRAM is becoming a precious resources, and the 3GB modules are being leveraged for AI chips but also, key products in Nvidia’s Mobile line.
Secondly, just in general – launching a new series of cards when GPU production is already kind of hammered is quite difficult.
But, the third and arguably biggest reason, AMD. Early reports for RDNA 4 (including myself) cited AMD were planning to launch the RX 9000 series with higher end SKUs, designed to compete with Nvidia’s flagship products. But, these were canned, and thus Navi 48 and 44 were the only chips which launched.
Nvidia’s flagships (such as the GeForce RTX 5080) simply face no pressure and competition, and AMD also isn’t releasing a refresh of any of its RDNA 4 parts this year, unless they’ve been very good at hiding the secret.
So, with AMD and Nvidia both apparently gearing up for a launch of their next generation parts (RTX 60 for Nvidia, and RDNA 5 or whatever it’s called for AMD)… well, yeah.
What does all of this mean for gamers? Well, ultimately the good news is that if you can pickup an RTX 50 or RDNA 4-based GPU right now, for a decent price, you’ll likely have the ‘best’ card you can for your money. On the other hand, if you’re happy with your current card and debating on waiting it out until next gen, you’ll have quite some time to save up.
