Get all your news in one place.
100’s of premium titles.
One app.
Start reading
Tom’s Hardware
Tom’s Hardware
Technology
Hassam Nasir

Nvidia RTX 5080 rumored to hit shelves on January 21 — expected to debut at CES with sales embargo reportedly lifting two weeks later

GeForce RTX 4090.

Rumor has it that Nvidia's RTX 5080 will hit shelves on January 21, says Hong Kong media outlet HKEPC (h/t VideoCardz). The source has a history of accurate information, but as always, we suggest you approach the leak with the appropriate skepticism. Additionally, the tipster has only revealed the embargo for the RTX 5080, so other GPUs in the Blackwell stack may launch on different dates.

Leaks suggest that Nvidia's Blackwell family will debut with the RTX 5090, RTX 5080, and RTX 5070 series. Nvidia is kicking off social media campaigns to drum up interest and has decided to revive GeForce LAN after 13 years. It is scheduled for 48 hours before Jensen's long-anticipated reveal on January 6.

HKEPC, in a concise and cryptic tweet, suggests that the RTX 5080 will retail starting from January 21. There's a nearly two-week gap between this date and January 6, when Nvidia is expected to launch the RTX 5080, a typical interlude for Nvidia GPU launches. If this rumor holds, we should see the RTX 5080 available at vendors on January 21. However, the leaker does not assert that other GPUs, such as the RTX 5090 and RTX 5070 series, could be delayed. In the past, Nvidia has employed a range of embargos for different GPUs in a specific family, just as we saw with the RTX 40 Super refresh.

The RTX 5080 allegedly features Nvidia's GB203-400-A1 at its core wielding 84 SMs or 10,752 CUDA cores. Akin to its predecessor, the RTX 5080 is outfitted with 16GB of memory though upgraded to much faster GDDR7 technology. Speaking of memory, it is rumored that apart from the RTX 5080 running 30 Gbps modules, the remaining RTX 50 series will stick with 28 Gbps memory. With a 10% CUDA core bump over the RTX 4080, expect most of the performance improvements to stem from architectural changes and the improved process node. Blackwell for servers (and likely desktop) is fabricated using a custom 4NP (5nm-class) node from TSMC, delivering 30% higher density than 4N used with Ada Lovelace.

Since AMD is not competing in the high-end space this generation, it appears the RTX 5080 will face minimal competition and remain largely uncontested. This allows Nvidia to dictate pricing without competition constraints. Still, we suggest you wait for the official keynote since pricing may be subject to last-minute adjustments.

Sign up to read this article
Read news from 100’s of titles, curated specifically for you.
Already a member? Sign in here
Related Stories
Top stories on inkl right now
One subscription that gives you access to news from hundreds of sites
Already a member? Sign in here
Our Picks
Fourteen days free
Download the app
One app. One membership.
100+ trusted global sources.