While the RTX 5080 has 360 W TDP, the NVIDIA GeForce RTX 5090 has 575 W TDP.

Kopite7kimi and Hongxing2020, two of the most reliable leakers, claim that NVIDIA’s GeForce RTX 5090 and RTX 5080 will have TDPs of 575 W and 360 W, respectively. According to earlier speculations, these GPU SKUs contain 600 W and 400 W TGPs, which stands for total graphics power. This means that a GPU, RAM, and other components use a specific amount of power. Thermal design power, or TDP, is a more precise figure that is linked to the GPU die or the particular SKU in question. Recent leaks indicate that the GeForce RTX 5090’s GB202-300-A1 GPU die uses 575 watts, while GDDR7 memory and other PCB components use 25 watts.

The GB203-400-A1 chip in the RTX 5080 is reportedly using 360 watts of power by itself, with an additional 40 watts allocated for GDDR7 memory and other PC components. Because its GDDR7 memory modules supposedly operate at 30 Gbps, the lower-end RTX 5080 consumes more power than the RTX 5090, which uses GDDR7 memory modules with 28 Gbps speeds. The first-generation GDDR7 memory may need more power to reach the 30 Gbps threshold, even though the RTX 5090 employs more modules or modules with larger capacities. More power is therefore allocated for that. It would be easy to increase speed without significantly increasing power in subsequent GDDR7 versions.