Nvidia’s new GPUs could be right around the corner

Jacob Roach / Digital Trends

Is Nvidia really about to add to its lineup of top GPUs? All signs point to yes, and now, we have an official Nvidia keynote on the horizon that tells us when we might hear more about the rumored RTX 40 Super. Nvidia revealed that it’s going to deliver a special address on January 8 as part of CES 2024. Although the company hasn’t confirmed what it’s planning to cover, the rumor mill has been buzzing with information about three new desktop GPUs. But will they really be worth the upgrade?

Several reputable leakers have weighed in on the matter of the RTX 40-series refresh, and we’ve been getting updates about the range for a few weeks now. Nvidia doesn’t need to specifically state that it’ll talk about these graphics cards, as that is going to be the expectation anyway. The three GPUs in question are the RTX 4080 Super, RTX 4070 Ti Super, and the RTX 4070 Super.

During its previous keynote at CES 2023, Nvidia talked about the RTX 40 laptop range but also unveiled the RTX 4070 Ti — formerly known as the RTX 4080 12GB. This year, as shared by VideoCardz, it’s likely that we’ll hear about all three new desktop GPUs, but some updates regarding AI are also possible. Although Nvidia might announce all three graphics cards at once, we don’t know whether all three will hit the market simultaneously. Leaks point to the first quarter of the year for all three, though.

Initially, it seemed like Nvidia might deliver a sizeable upgrade with the RTX 40-series refresh. Sources claimed that Nvidia might shake things up, upgrading the RTX 4080 Super to the AD102 GPU (which is the flagship chip found in the RTX 4090). The RTX 4070 Super was said to be getting upgraded to the AD103 chip, which is currently utilized by the RTX 4080. These changes would unlock far more cores, an increase in VRAM, and a larger memory bus for both cards.

MSI RTX 4080 Suprim X installed in a PC.
Jacob Roach / Digital Trends

However, it’s best not to put much stock in these assumptions. Other leaks suggest that the RTX 40 refresh will be a minor upgrade, with all cards sticking to their respective chips and maxing out the remaining core counts. Some rumors pin the RTX 4080 Super as featuring 10,240 CUDA cores and 16GB of GDDR6X memory across a 256-bit bus. The RTX 4070 Ti Super is said to sport 8,448 CUDA cores and 16GB VRAM, while the RTX 4070 Super might have 7,168 cores and 12GB VRAM.

Those specs mark an increase over the non-Super cards, but not a massive one. The RTX 4070 Ti Super (as a side note, that name’s always going to be a mouthful) would see the greatest boost thanks to the jump from 12GB to 16GB VRAM, while the core count boosts are mostly minor.

Regardless of the specs, the RTX 40 Super lineup might serve one important purpose — it should help the prices of some GPUs go down. Nvidia’s pricing has been on a worrying upward trend, which doesn’t bode well for the next generation of GPUs. If we can see the existing RTX 4070, RTX 4070 Ti, and the RTX 4080 drop in price thanks to the refresh, that alone will be a good thing.

We’ll be covering CES 2024 live as it happens, so stay tuned to make sure you don’t miss Nvidia’s announcement.

Editors’ Recommendations






This is the cheapest GPU that’s still worth buying

In a market that’s centered around all of the best graphics cards, it can be deceptively hard to buy a GPU. For those of us who don’t need “the best” and just want “good enough,” it’s a tricky process. Sometimes spending a few more dollars gets you something that’s considerably better, and other times, spending $100 less means a GPU with significantly better value.

Fortunately, there are a few GPUs that are just all-around good picks right now. If you want to get the most GPU power for the least possible money, check out the AMD Radeon RX 6600.

AMD Radeon RX 6600

Jacob Roach / Digital Trends

On the whole, AMD is often cheaper than Nvidia, and that applies both to last-gen and current-gen cards. It has a few noteworthy cards in the $180 to $320 range, not to mention the $500 RX 7800 XT, which is one of the best-value graphics cards in this generation. However, if you want to get the cheapest GPU possible while still retaining decent performance, I recommend the RX 6600.

The RX 6600 belongs to AMD’s RDNA 2 lineup, and it has two slightly more powerful siblings — the RX 6600 XT and the RX 6650 XT. However, due to their similar performance, you might as well go for the cheapest one when you’re on a tighter budget. The RX 6600 can be found for as little as

, and at that price, it’s a good deal.

What kind of performance can you expect from this card? You should be able to play most titles at 1080p without needing to compromise on the settings too much. In our own testing, the RX 6600 was able to tackle Red Dead Redemption 2 and maintain 59 frames per second (fps) on average — and that’s at ultra settings. We’ve also tried it in Fortnite, Assassin’s Creed Valhalla, Battlefield V, and Civilization VI, and it averaged between 71 fps and 138 fps. Read AMD Radeon RX 6600 review to learn more.

While I recommend this GPU, it’s not a straightforward choice, because AMD has a slew of RX 6000 cards that are in a similar price range and offer slightly better performance. There’s the RX 6600 XT, RX 6650 XT, and the current-gen RX 7600. The latter can be found for around

, and if your budget can stretch that far, it’s money well spent.

When we benchmarked the RX 7600, we found that it averaged 66.3 fps across our 1080p test suite, while the RX 6600 scored 50.9 fps. If you want a steady 60 fps in modern games at 1080p, the RX 7600 is a more straightforward pick; if you don’t mind dropping down to medium settings, go for the RX 6600.

What about Nvidia and Intel?

Two intel Arc graphics cards on a pink background.
Jacob Roach / Digital Trends

If you’re not that into AMD, but still want a cheap GPU, there are a couple of options from Intel and Nvidia that are worth considering: the Intel Arc A750 and the Nvidia RTX 4060.

One thing to be mindful of when buying Intel Arc cards is compatibility. As you can read in our in-depth review of the Arc A750 and the Arc A770, while these GPUs are compatible with older processors, they benefit from the ability to enable Resizable BAR. That means you’ll need a 10th-gen Intel processor or an AMD Ryzen 3000 or newer. Without Resizable BAR, the performance of Arc A750 drops by a large margin.

With that disclaimer out of the way, the Arc A750 is a surprisingly solid GPU for 1080p gameplay, averaging 80 fps at 1080p ultra across our entire test suite. It’s also viable for 1440p, hitting 62 fps on average, while the Arc A770 scored 68 fps.

If you want to try out Intel in your next PC build, the cheapest version of the Arc A750 costs

and comes from Sparkle. Intel’s own version is slightly pricier at $230.

Moving on to Nvidia, as you can read in our review of the RTX 4060, this card comes with a few caveats, but it still has an edge over the competition in that it has access to Nvidia’s Deep Learning Super Sampling (DLSS 3) frame generation. While not supported in all too many games, DLSS 3 can help an RTX 4070 outperform an RTX 4090, and it has a similar effect on the RTX 4060. It’s an artificial boost that opens the door to high-end gaming on a tight budget, which is why the RTX 4060 makes sense despite the price tag.

Without the DLSS 3 crutch, the RTX 4060 still performs well enough to handle 1080p gaming with ease. It averaged 78.2 fps in our 1080p test suite at ultra settings. As for DLSS 3, it really does make a difference on a GPU of this caliber — while the 4060 hits 82 fps on average without DLSS 3 in Cyberpunk 2077, it managed a whopping 136 fps with frame generation enabled.

The RTX 4060 starts at around

.

Do you really need a GPU?

AMD Ryzen 5 5600G retail box with processor on white background.
AMD

If you’re trying to build a PC while spending as little as possible, and you’re not a hardcore gamer, one way to save a couple hundred dollars is to not get a graphics card at all.

Integrated graphics often get a bad rep, and rightfully so, but we’ve come a long way in the past few years. In particular, one option that’s viable for daily use (including some gaming) is AMD’s Ryzen 5600G processor, or rather, accelerated processing unit (APU). This is a separate category of processors that combine the CPU and the GPU within a single chip. This integration gives the 5600G more GPU power than most iGPUs.

I’m not going to lie to you — the Ryzen 5600G is no gaming beast. However, as long as you’re content with medium settings and you’re mostly playing older games and indies, it’ll do. It’s also a solid option for day-to-day tasks, be it work or casual browsing. Most of all, it’s highly affordable — it’s priced at around

right now. There’s also the Ryzen 7 5700G, with more cores and slightly better performance. It’s priced at

.

One good way to make the most of this APU is to ensure that you’re using some of the best RAM available in your price range. This is because the Ryzen 5600G doesn’t have its own dedicated video memory (VRAM), which a discrete GPU would normally have. To make up for that, it uses a portion of your system’s RAM to store textures and other graphics-related data. The faster your RAM, the better the performance.

No matter the GPU or APU you pick, don’t be afraid to dip into the previous generation when assembling a budget PC. Some of those options are still pretty excellent and, following several discounts, now offer far more agreeable prices than some of their current-gen counterparts.

Editors’ Recommendations