When there are just over 2 months left until NVIDIA presents the new range of graphics RTX50an event scheduled for the first week of January at the CES held in Las Vegas, we already know almost all the specifications of the RTX 5090 and RTX 5080, the two graphics cards that NVIDIA will announce first.
Based on these specifications, the TGP that the RTX 5090 will have will be 600W. We learned about this information a few weeks ago and it gave rise to a rumor that NVIDIA’s highest-end graphics card would use two power connectors 16-pin, something that is partly logical seeing the problems that the RTX 4090 had with the power cable.
However, it seems that this will not be the case in the end. The power supply manufacturer Segotep has confirmed through the Chinese social network Bilibili that this is not the case. This claims that it has the RTX 5090 specifications and denies the rumors that this graphics card incorporates two power connectors.
PC component manufacturers know in advance the specifications of the components that are launched on the market in order to satisfy their needs in terms of performance, so, initially, although it is a power supply manufacturer that does not have presence outside China, we could take this information as true.
In this way, it is confirmed that the substitute for the RTX 4090 on the market, it will use the same 16-pin power connector. In this same publication, Segotep also states that we will see you in January, confirming that the RTX 5090 will be presented as planned in the absence of official confirmation.
ATX 3.x power supplies have one 12V-2×6 connector and three 8-pin PCIe connectors, so if NVIDIA had gone down the route of using two power connectors, it would have been a problem for users who had the intention to buy it when forced to invest in a new power supply.
What do we expect from the NVIDIA RTX 50
After many months of rumors and speculation, it seems that we finally know what specifications NVIDIA’s top-of-the-range graphics card will have, a graphics card managed by the company. GPU202-300-A1 with 21,760 CUDA Cores. As for memory, we are talking about 32 GB of memory GDDR7 with 512-bit memory interface.
Regarding the RTX 5080, it will have a configuration similar to the RTX 4080, with 16 GB of video memory, but 256-bit GDDR7 type and will have a consumption of 400W. This model is based on the GPU 203-400-A1.
AMD announced at the beginning of August that it was abandoning the graphics race against NVIDIA to fight for the high-end, a decision that will probably be temporary and that, we hope, will not last long. For now, the RX Series 8000 It is scheduled to be launched on the market during the first quarter of 2025.
Regarding Intel with its second generation of dedicated graphics, we know little or nothing, although the plan announced by the company is to present them before the end of 2024.
