In today's announcements at CES 2024 here in Las Vegas, Nvidia officially unveiled its RTX 40 Super Series line up of GPUs, with the 4070, 4070 Ti, and the 4080 getting the treatment. All three GPUs are set to launch this month.
These new GPUs aim to provide a performance bump over the non-super variants Nvidia released previously, with the line up seeing higher CUDA core counts and, in some cases, more and faster memory.
RTX 4080 Super
This one is the most interesting to me, as it sees the RTX 4080-class of GPU able to compete price-wise with rival AMD's flagship, the RX 7900 XTX. While both cards traded blows in our review, with the Nvidia card proving better with ray tracing applications specifically, the price gulf between the two made the Nvidia card a little tougher pill to swallow.
Now, with the RTX 4080 Super hitting the line up Nvidia GPUs at the same $999 price tag (replacing the old 4080 as well in the stack), the pressure is on AMD, especially when you factor in more robust support for DLSS 3 frame generation on Nvidia's hardware versus AMD's Fidelity FX Super Resolution 3's frame generation - which only has a few games that have adopted it thus far.
The RTX 4080 Super is the full AD103 chip, with more CUDA cores than its predecessor, as well as faster memory bandwidth. This feels like it's targeted at those who've yet to upgrade from an RTX 2080 or 3080 Ti while pricing more competitively with AMD's similarly performing GPUs. The RTX 4080 Super launches on January 31st.
RTX 4070 Ti Super
The RTX 4070 Ti Super, meanwhile, stays at its $799 price range as the original variant launched, taking over its spot in Nvidia's GPU stack. The big news for the RTX 4070 Ti is the increase in memory to 16GB GDDR6X as well as a bump to the 256-bit memory bus, but also an increased CUDA core count over the card it's replacing.
If you'll recall, the RTX 4070 Ti was originally meant to be tagged as the RTX 4080 12GB, though after consumer backlash Nvidia walked this back and rebranded it as the 4070 Ti GPU. As a result, there are no Founder's Boards for this class of card, and the same holds true for the RTX 4070 Ti Super. This will launch on January 24th and, if Nvidia's internals bear out in real-world testing, there should be about a 10% performance bump over the existing RTX 4070 Ti. Naturally, wait for independent reviews that will take a look at that claim.
RTX 4070 Super
And finally, the RTX 4070 Super. This is the only card announced today that is not replacing its predecessor in the GPU stack. Instead, the price drop Nvidia gave the RTX 4070 late last year is staying with the older model, while the Super variant slides into the stack at that $599 price tag. Nvidia's big claim with the RTX 4070 Super is that it will be faster than the RTX 3090 - the last generation's flagship GPU. While in our pre-brief, they did acknowledge that in some games, the two cards will trade blows with the 3090 coming out ahead, on average, the RTX 4070 Super should be faster - and that's before you turn on frame generation. With frame gen, the card outperforms the previous generation, at least according to Nvidia, hands down.
The RTX 4070 Super is still sitting on the same 192-bit memory bus as its RTX 4070 non-super brother, Nvidia is boosting the power draw by about 20 watts to eke out the extra performance the Super can provide. Launching on January 16th at $599, the RTX 4070 Super looks to be a good little mid-range GPU, should the claims of faster than 3090 performance and an Nvidia-calculated 15% performance boost over the original 4070 hold true.
