We know that it’s in the next few hours, at CES 2025 on January 6, that CEO Jensen Huang will present the new “Blackwell” architecture and several RTX 50s from this new range. So far, several hypotheses have emerged concerning the commercial launch of the cards after this event. HKEPC reported that the GeForce RTX 5080 will be launched on January 21, 2025, then withdrew its message from its social networks. The RTX 5090 is expected to follow suit and hit the market before the end of January, according to the rumors we’ve heard so far.
RTX 5080, Intel B570, RDNA 4: A busy January
For the moment, it’s hard to say where these RTX 50s stand. We know from numerous leaks that the RTX 5080 will use a GB203-400 graphics chip. This will give access to a configuration of 10,752 CUDA cores, accompanied by 16 GB of 30 Gbps GDDR7 memory which, with a 256-bit memory interface, will deliver a bandwidth of 960 GB/s. Power consumption is expected to be at least 400W, justifying the use of a new 12V-2×6 16-pin connector. How this future RTX 5080 compares with the current top-of-the-range RTX 4090 is objectively difficult to say. The RTX 4090’s datasheet still seems superior to what the 5080 is expected to offer. However, the latter’s supposedly higher power consumption suggests that Nvidia will be revising its operating frequencies upwards in order to get closer to it.
However, one thing is clear: the January calendar will be extremely busy. Intel is expected to lead the way with the arrival of the B570, followed by Nvidia’s launch and then AMD’s RDNA 4 cards. It’s quite possible that AMD’s schedule will be re-evaluated in the light of Nvidia’s announcements at the start of CES.
