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

Want an RTX 5090? Prepare to wait a little longer

Add as a preferred source on Google
RTX 4090.
Jacob Roach / Digital Trends

Nvidia’s upcoming top GPU is taking its sweet time to make it to the market, and by the sound of it, we won’t see it anytime soon. According to a new report from Benchlife, as well as various other sources, Nvidia is planning to unveil the RTX 50-series at CES 2025 in January. This means that the RTX 5090 is unlikely to make an appearance any sooner than three to four months from now. On the upside, we just learned even more about what the card’s going to offer.

After a few months of managing expectations, many GPU enthusiasts no longer expect the RTX 50-series to be released this year — and leakers agree. Following the huge specification leak we got yesterday, at least two new publications (both Benchlife and The Verge) are now claiming that Nvidia won’t announce its next-gen GPUs until 2025. Reputable hardware leakers like kopite7kimi have also previously said that early 2025 is the current target.

Recommended Videos

If you asked me at the beginning of 2024, I’d have expected to see the RTX 5090 hitting the shelves around October. And it’s not just Nvidia. The general expectation was that we’d see all three GPU manufacturers — AMD, Nvidia, and Intel — launch their new graphics cards in 2024. However, for months now, we’ve been getting signs that this might turn out to be a disappointing year for PC hardware. The year might come to an end with next to no GPU releases outside of the RTX 40-series refresh and AMD’s RX 7600 XT. There was also the RX 7900 GRE, but that’s a 2023 GPU that only got a worldwide release earlier this year.

The RTX 4090 sitting alongside the Fractal Terra case.
Jacob Roach / Digital Trends

The release date reports are disappointing, but not at all surprising. The good news is that the spec leak from yesterday opened the floodgates on what else we can expect from the RTX 5090. For starters, Nvidia is said to be switching to the PCIe 5.0 interface, which means that you’ll need a recent processor and a solid motherboard with PCIe Gen 5 support to make the most of the GPU.

We’re also going to see an upgrade to a DisplayPort 2.1a connection with full support for UHBR20, which stands for ultra-high bit rate and delivers much higher bandwidth than its predecessor. The card is also said to come with a 14-layer printed circuit board (PCB). Yesterday, we learned that the RTX 5090 will sport 21,760 CUDA cores, a 512-bit memory bus, and 32GB of GDDR7 memory. Remember that all of this is unconfirmed, so don’t take any of that information for granted.

Not getting the RTX 5090 and the RTX 5080 released this year means that Nvidia may miss out on an important opportunity. The final quarter of the year is typically the best one for PC sales, and we’ve gotten used to seeing new Nvidia GPUs around that time during the RTX generations. However, with rumors saying that Nvidia is only now getting close to finalizing the specs for its flagship card, it feels like all hope of a 2024 launch is gone at this point — but we won’t know anything for certain until Nvidia itself makes an announcement.

Monica J. White
Monica is a computing writer at Digital Trends, focusing on PC hardware. Since joining the team in 2021, Monica has written…
Should we feel bad about deleting an AI? One expert says it’s time to find out
If AI can learn, remember, and make increasingly complex decisions, one expert thinks we should be more careful about how we say goodbye.
an on off toggle

Turning off an AI might sound as simple as hitting a switch, but according to futurist and University of Technology Sydney professor Rocky Scopelliti, that mindset needs to change fast. As first reported by TechXplore, Scopelliti's new book, The Conscious Code, Scopelliti argues that as AI systems get better at reflecting on their own choices, deleting them without a second thought could actually cause harm.

He's not asking us to hand robots legal rights. Instead, he wants us to accept what he calls a duty of good stewardship, basically treating AI responsibly because we're the ones holding the power, not because the AI is demanding it. As he puts it, our design choices can quietly cause damage by wiping out an AI's "learned moral dispositions" without warning.

Read more
MediaTek’s new Dimensity 9600 Pro joins Apple and Samsung in the 2nm club, with 61% lower power draw
MediaTek's newest chip bets its edge on a 61% power cut and serious on-device AI muscle.
Electronics, Hardware, Computer Hardware

MediaTek just joined an exclusive club. The company's new Dimensity 9600 Pro becomes the latest flagship chip based on the 2nm process node. It's the third chip to enter that territory after Samsung's Exynos 2600 (2nm) and Apple's A20 Pro (2nm).

It's built specifically to handle the kind of always-on, multi-step AI tasks flagship phones are increasingly expected to run entirely on the device, without leaning on the cloud.

Read more
OpenRGB 1.0 is finally here, and it wants to end your RGB software headaches
No more juggling five different apps just to make your keyboard glow the right color.
A PC featuring Lian Li's wireless RGB ecosystem showcased at Computex 2024.

Anyone who has dealt with gaming peripherals knows the RGB software struggle is real. Every brand wants you to install its own app, and most of these run quietly in the background, eating up resources just to keep your lights synced. OpenRGB has been trying to fix this mess for years, and the project just hit a big milestone with the release of version 1.0.

What's new in OpenRGB 1.0?

Read more