Corsair just spilled the beans on next-gen GPU requirements

1 week ago 3
Nvidia GeForce RTX 4090 is shown along with a hand holding the power cable adapter. Jacob Roach / Digital Trends

Sometimes, news about next-gen GPUs comes from unlikely sources — today is one of those days. Corsair just spoke about its power supply units (PSUs) and cooling solutions in relation to the future of some of the upcoming best graphics cards. It turns out that Nvidia’s RTX 50-series may not be that much more power-hungry than the current-gen cards, but there’s more than just Nvidia to consider here.

Although unexpected, Corsair’s statement sounds like good news. The company doesn’t talk about any new solutions. In fact, Corsair seems to confirm that the PSUs we use today will still work fine for next-gen cards — provided the wattage is sufficient.

“The latest graphics cards now require a 12V-2×6 GPU power connector, a standard we expect will continue with next-gen GPUs,” said Corsair.

Get your weekly teardown of the tech behind PC gaming

On the other hand, the company admits that there could be a rise in power requirements: “Next-generation high-end graphics cards could demand even more power than current models, which can draw up to 450 watts. Factoring in high-end CPUs and other power-hungry components, this points to an increased need for higher wattage PSUs.”

This means that we could see a more power-hungry RTX 5090, but still, the connector can only handle up to 600 watts — so that’s the limit.

Corsair's power connectors.Corsair

There’s been a lot of talk about the power requirements for next-gen GPUs, although that may be too broad a statement. In reality, the speculation is mostly centered around Nvidia, namely its flagship RTX 5090. AMD and Intel, while also set to launch RDNA 4 and Battlemage, are unlikely to release enthusiast-level cards that need serious cooling and consume lots of power. Nvidia, however, is almost certainly pushing for the high-end.

We’ve seen claims that the RTX 5090 will require 600 watts of power — that’s 150W more than its predecessor. We’ve also seen reputable leakers say that the entire lineup will see an increase in power consumption. However, some of the claims were a bit too wild to be real, such as the speculation that the RTX 5090 might use not one, but two 12V-2×6 power connectors.

Corsair doesn’t seem to imply that any of these outlandish claims are true, which is good news. If Nvidia sticks to the 12V-2×6 power connector, then we’re guaranteed to have the RTX 5090 with a power requirement of under 600 watts. Although Corsair doesn’t mention Nvidia by name, it’s unlikely that either AMD or Intel will have cards that need nearly as much power. However, they might still use the same power connector — we’ll have to wait and see.

Monica J. White

Monica is a computing writer at Digital Trends, focusing on PC hardware. Since joining the team in 2021, Monica has written…

How to know which Mac to buy — and when to buy it

The M4 Mac mini being used in a workplace.

If you’re in the market for a new Mac (or Apple display), there’s a lot of choice ahead of you. Maybe you're interested in a lightweight MacBook Air from the selection of the best MacBooks -- or maybe one of the desktop Macs. Either way, there’s a wide variety of Apple products on offer, including some external desktop monitors.

Below you'll find the latest information on each model, including if it's a good time to buy and when the next one up is coming.
MacBook Pro

Read more

AMD Ryzen AI claimed to offer ‘up to 75% faster gaming’ than Intel

A render of the new Ryzen AI 300 chip on a gradient background.

AMD has just unveiled some internal benchmarks of its Ryzen AI 9 HX 370 processor. Although it's been a few months since the release of the Ryzen AI 300 series, AMD now compares its CPU to Intel's Lunar Lake, and the benchmarks are highly favorable for AMD's best processor for thin-and-light laptops. Let's check them out.

For starters, AMD compared the Ryzen AI 9 HX 370 to the Intel Core Ultra 7 258V. The AMD CPU comes with 12 cores (four Zen 5 and eight Zen 5c cores) and 24 threads, as well as 36MB of combined cache. The maximum clock speed tops out at 5.1GHz, and the CPU offers a configurable thermal design power (TDP) ranging from 15 watts to 54W. Meanwhile, the Intel chip sports eight cores (four performance cores and four efficiency cores), eight threads, a max frequency of 4.8GHz, 12MB of cache, and a TDP ranging from 17W to 37W. Both come with a neural processing unit (NPU), and AMD scores a win here too, as its NPU provides 50 trillion operations per second (TOPS), while Intel's sits at 47 TOPS. It's a small difference, though.

Read more

This fps-doubling app is now even better than DLSS 3

Cyberpunk 2077 on the Sony InZone M10S.

Lossless Scaling is a $7 Steam app that's flipped the idea of frame generation on its head this year. Similar to tools like Nvidia's DLSS 3 and AMD's FSR 3, Lossless Scaling offers frame generation. However, it works with any game, and with any graphics card, and it can triple or quadruple your frame rate with this frame generation. And now, the app is going further with a feature that even DLSS 3 and FSR 3 don't have.

The developer posted the 2.12 beta to Steam on Wednesday, and it adds a couple of new features. The big one is a resolution scale for LSFG, the tool's own machine learning-based frame generation algorithm. This allows you to decrease the resolution of the input frames, leading to a very minor quality loss in exchange for a fairly large performance boost. The resolution of the game doesn't change at all. You're basically giving the frame generation algorithm slightly less information to work with.

Read more

Read Entire Article