Nvidia on Friday announced it is “pressing the ‘unlaunch’ button” on the RTX 4080 12GB, one of the three “Ada Lovelace” 40-series graphics cards it unveiled last month.
“Having two GPUs with the 4080 designation is confusing,” the company said in a blog post announcing the move, which verged on the edge of satire.
The RTX 4080 12GB garnered quite a bit of backlash from the tech community because its name is the only thing it shares with its older sibling, the RTX 4080 16GB. It’s not just about the RTX 4080 16GB having four extra gigs of GDDR6X VRAM — in reality, the “4080 12GB” doesn’t even have the same GPU chip as the 4080 16GB.
The 4080 16GB features the AD103 chip, Nvidia’s second-largest for this generation, while the 4080 12GB has the smaller and less powerful AD104 chip under the hood. As such, the 4080 12GB has significantly fewer CUDA cores (7,680 vs. 9,728) and, not to mention, a narrower memory bus (192-bit vs. 256-bit).
What’s more, the RTX 4080 12GB is on average only 12% faster than the card it was supposedly positioned to succeed, the RTX 3080. That does not look like a generational improvement. Nvidia’s naming scheme for the two RTX 4080 cards was confusing at best and misleading at worst.
Nvidia “unlaunching” the 4080 12GB is good.
BUT
The 4080 is still selling for a ridiculously absurd $1200.
And they’ll just call the 4080 12GB a 4075 so everyone stops complaining.https://t.co/8vGJkvsV3x— Brett @ UFD Tech (@UFDTech) October 14, 2022
In all likelihood, Nvidia will simply rebrand and relaunch the RTX 4080 12GB as a lower-tier RTX 4070 (or similar) down the line.
Meanwhile, the RTX 4080 16GB (or should we simply call it the RTX 4080 now?) is still on track to launch next month on November 16. It is 30% faster than the now-“unlaunched” 4080 12GB and actually looks to be worthy of the “4080” moniker.
Nvidia’s top-of-the-line halo card from the 40-series, the RTX 4090, hit shelves earlier this week on October 12.
