News

NVIDIA GeForce RTX 5080: specifications, price and everything we know

Virtually all hardware fans are going to be very attentive to CES 2025, which will take place in Las Vegas at the beginning of January, since among other things it is the stage in which supposedly NVIDIA will officially present its new generation of graphics. Meanwhile, and as usual, rumors and leaks have continued to give many of the models a practically complete shape, so today we are going to concentrate everything we know until now of the RTX 5080surely one of the most anticipated models of the new generation of those in green.

Based on the Blackwell architecture, the RTX 5080 will be the second most powerful graphics card of the new generation, and will replace the RTX 4080 and RTX 4080 SUPER of the current generation. It will of course be a fairly expensive graphics card, most likely above 1,000 euros, but it will also be the “sweet spot” for those who want to build a high-performance gaming PC. Let’s go see it.

RTX 5080, “confirmed” technical specifications

The RTX 5080 with Blackwell architecture will be manufactured on a 5nm node, more specifically on TSMC’s N5B. The chip used will have the name GB203-300 A1successor to the AD103-300 A1 of the RTX 4080, and will be mounted on the PG144/147-SKU45 PCB with 84 SMs. This is all the confirmed technical information about the die, but we can still go a little further if we rely on rumors…

The NVIDIA GeForce RTX 5080 would have 10,752 CUDA cores with 84 SMs, which represents an increase of around 10.5% compared to the previous generation. As for the rest of the technical specifications of the GPU, including SM, Tensor Cores, RT Cores and operating speed, there are only assumptions without solid foundations, so for now we are going to consider that they are still unknowns.

NVIDIA is expected to maintain the same amount of memory, 16 GB of VRAM (although there are rumors that point to a variant with 24 GB), but it would be upgraded to the new generation GDDR7 compared to the current generation GDDR6X. This memory would operate on a 256-bit bus with an effective speed of 32 Gbps, which would give us a memory bandwidth of 1 TB/s, exceeding the 1,007 GB/s of the current RTX 4090.

In terms of consumption, it is assumed that the RTX 5080 will be higher than the RTX 4080, and if it has 320W of TBP it is quite safe to expect a consumption of around 400W. Remember, however, that the TBP is the consumption value that the graph as a whole can reach, but it is not necessarily its “normal” consumption in gaming. In any case, with a TBP of 400W we can almost guarantee that NVIDIA will repeat the use of the 12VHPWR connector of the previous generation, which as you well know is capable of supplying up to 600W of power with a single cable.

12VHPWR

Finally, in terms of performance This new graphics card is expected to offer around 20-25% more than the RTX 4080 SUPER, and in fact many experts suggest that it would offer 10% better performance than the current RTX 4090. There’s not much more to say about it until we have the graph in our hands to test it.

Price and availability

The two most controversial aspects, as always. We’ve already mentioned that NVIDIA is expected to make the official announcement at CES in January, and that could mean in-store availability could be towards the end of January, if we go by precedent.

As for the price, we can only make assumptions. The RTX 4080 was launched with an official price of $1,199, but when the RTX 4080 SUPER arrived, its price dropped to $1,000. Therefore, we personally expect the RTX 5080 to cost exactly the same, which would translate to about 1,200 euros in Spain, perhaps more depending on the assembler and the custom model they put on store shelves.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *