Currently, the RTX 5090 serves as the flagship model in NVIDIA’s consumer graphics card lineup. It holds the title of the world’s most powerful GPU and is also the most expensive consumer graphics card. Finally, there’s a significant gap in price and performance compared to the card just below it in the manufacturer’s lineup: the RTX 5080. While a 5090 retails for €4,000, the RTX 5080 is available for between €1,300 and €1,400. And that’s where the rumor of an RTX 5090 SE—a mid-range option in terms of performance and price—makes sense!
RTX 5090 SE: A mid-range card between the RTX 5080 and the RTX 5090!
Still just a rumor!
Broadly speaking, this card would feature a partially disabled GPU, as it would have only 110 SMs, for a total of 14,080 CUDA cores. In contrast, the original 5090 features a configuration with 170 SMs and 21,720 CUDA cores—nearly 55% more. Furthermore, the RTX 5080 features 84 SMs and 10,752 CUDA cores—23% fewer than this hypothetical SE card.
As for memory, we’d see a memory bus that falls between the two cards, with 384 bits and 32 GB ofGDDR7. As a reminder, the 5090 has a 512-bit bus, compared to 256 bits for the 5080. In terms of clock speeds, theGDDR7 chips are expected to operate at the same speed as the original 5090: 28 Gbps.
In terms of power consumption, a TGP of 500W is being discussed, while the original 5090 runs at 575W by default.
A price of $1,500—is there still hope?
Finally, if we’re to believe what our colleagues are saying, NVIDIA could be targeting a retail price of $1,500. Keep in mind, this would be the MSRP; in reality, the actual cost would be much higher than that.
Of course, these are just rumors, and this card would likely replace a hypothetical RTX 5080 Ti. In contrast, the lower segment—currently occupied by the RTX 5080—could be updated with the RTX 5080 Super, which is expected to offer increased VRAM capacity and potentially higher clock speeds.
