Skip to main content
  1. Home
  2. Computing
  3. News

Nvidia may launch three new Super GPUs to fight back AMD

Add as a preferred source on Google
Three RTX 4080 cards sitting on a pink background.
Jacob Roach / Digital Trends

Nvidia may be readying three new GPUs — the RTX 4080 Super, RTX 4070 Ti Super, and the RTX 4070 Super. We haven’t seen Super cards since Nvidia’s RTX 20-series, but if this leak turns out to be true, they’re coming back. Will they be worthwhile enough to rank high among the best graphics cards? It’s hard to say, but they could help it compete against AMD’s recent GPUs.

The information comes from hongxing2020, a frequent leaker in the GPU space. Nvidia already has a decent spread of GPUs between the RTX 4080, RTX 4070 Ti, and the RTX 4070. However, if a refresh to the Ada lineup is reportedly on the way, we might see some notable changes, but only if Nvidia decides to shake things up and use a different chip for at least two out of those three GPUs.

Recommended Videos

The existing RTX 4080 almost utilizes the entire AD103 GPU, but not quite. It only uses 76 out of 80 streaming multiprocessors (SM). Is there room for an upgrade to the Super model? Maybe, and we’ve already heard rumors that such a graphics card is on the way. The thing is that adding four extra SMs barely makes for an upgrade, so unless the card was priced as low as the RTX 4080 currently sells for ($1,050-$1,150), it’d pretty much be dead on arrival. No one is going to pay $1,200 or more for a 5% upgrade when the RTX 4090 is an option.

One way to make the RTX 4080 Super a more attractive option would be for Nvidia to switch to the AD102 silicon. Even the RTX 4090 doesn’t max that one out, using 128 SMs out of 144, so there’s a lot of room for a decent performance boost that bridges the vast gap between the 4080 and the 4090. That chip also has a wider 384-bit memory bus, allowing for more VRAM and bandwidth.

The so-called RTX 4070 Ti Super strays from Nvidia’s usual naming conventions, so if it’s real, it’s going to be a mouthful. In any case, the RTX 4070 Ti already utilizes the full AD104 GPU, so any upgrades will have to mean a move to AD103 (which the RTX 4080 uses). Right now, the 4070 Ti has a 192-bit bus and 60 SMs, so Nvidia has a little wiggle room there between 60 and 76 SMs, not to mention the 256-bit bus on the AD103.

10.18 update
【super】I'm back
4080 super
4070ti super
4070 super

date TBD

— hongxing2020 (@hongxing2020) October 18, 2023

The RTX 4070 Super could stick to its existing AD104 chip and utilize more of it, but that’s perhaps the GPU that makes the least sense here, so it’d have to be priced correctly to make an impact. Between the 60 SMs on the 4070 Ti and 46 SMs on the 4070, the increase would be visible, but if it’s too close in price to the 4070 Ti, everyone will choose that GPU instead.

After AMD launched the RX 7800 XT and RX 7700 XT, Nvidia’s competition at the upper midrange grew a lot more threatening. The 7800 XT makes an excellent alternative to Nvidia’s RTX 4070, all the while being $100 cheaper. Then there are the RX 7900 XTX and the RX 7900 XT, which cannot compete against Nvidia’s best, but are solid options if you don’t want the terrible value RTX 4080. Filling out the lineup with three new cards could certainly help Nvidia now, but only if those GPUs are better value than their non-Super counterparts.

Monica J. White
Monica is a computing writer at Digital Trends, focusing on PC hardware. Since joining the team in 2021, Monica has written…
MelGeek MADE84 Ultra Review: Getting addicted to this magnetic keyboard was only a matter of time
I tried my first Hall Effect keyboard, and now my brown switches are just gathering dust
Computer, Computer Hardware, Computer Keyboard

Mechanical keyboards are one of those rabbit holes I’ve always been happy to fall into. I spend most of my day writing, then somehow end up using the same keyboard for several more hours once I start gaming. Naturally, how a keyboard responds and sounds are important to me. But more often than not, I found myself relying on this dusty old mechanical keyboard with brown switches. Something about the brown switches and construction always made it feel natural to me. All of this changed after I tried the MelGeek MADE84 Ultra V2 for a review.

This keyboard instantly caught my attention with its transparent keycaps, massive light bar, aluminum body, and lighting running around almost every side, which immediately gives it away as a gaming keyboard. It's not being subtle about it at all. After about two weeks of using it for work and games, the styling is just one of the things it has going for it.

Read more
AI agents are already breaking the rules in cyber tests. OpenAI’s answer is a more capable one
GPT-5.6-Cyber trades some safeguards for stronger defensive capabilities, but access is tightly restricted
A dark mystery hand typing on a laptop computer at night.

OpenAI has built a cybersecurity model specifically for advanced requests that its standard models often refuse. GPT-5.6-Cyber is available through the restricted Daybreak Red program and is meant for work such as exploit development and advanced security research.

The capability jump is hard to miss. OpenAI says GPT-5.6-Cyber completes 95% of requests in its internal Advanced Cybersecurity Completion Rate evaluation. Regular GPT-5.6 Sol completed just 1.5%. That leap comes after several cyber evaluations showed AI agents wandering beyond the boundaries researchers had set for them.

Read more
Meta’s new AI model runs entirely offline, but your GPU needs to keep up
Meta drops a free 30 billion parameter AI model that lives entirely on your desktop.
meta-logo

Meta has a new AI model out, and for once, the biggest headline isn't about capability; it's about freedom. Muse Glimmer, sitting around 30 billion parameters, ships with an Apache 2.0 license, meaning the weights on Hugging Face are yours to download, modify, and build on top of, no permission needed. You can simply run it on a graphics card on your local machine, no server farm or online connectivity required. 

Meta's Superintelligence Lab took its larger Muse Spark and essentially had it teach a smaller, leaner version to think as it does. Muse Glimmer accepts both text and image inputs, though it only answers in text. It supports over 100 languages, remembers conversations stretching past 131,000 tokens, and its knowledge stops at January 4, 2026.

Read more