NVIDIA’s next-gen Blackwell graphics cards will have roughly the same TBP (Total Board Power) as the RTX 40 “Ada Lovelace” family. According to Benchlife’s sources, NVIDIA is testing and validating heatsinks for the RTX 5080 and 5090. Four designs are being evaluated, ranging from 250W to 600W solutions. This coincides with the heatsink designs developed for the RTX 4090 (and the canceled 4090 Ti) which tops out at 450W.
Manufacturers usually test a wider range of cooling solutions than needed. We saw this before the RTX 4090 release where rumors claimed 600W+ designs, but none materialized. Even if they existed, these weren’t meant for the consumer market. The same should be the case with Blackwell. NVIDIA and its partners may test heatsinks built for dissipating 600W chips, but the final round of verification/validation will only include lower TBP variants (<500W).
Seeing as the GeForce RTX 4080 Super has a TBP of 320W, it’s likely that the RTX 5080 will also have a power budget of ~250-350W. A TBP of less than 300W shouldn’t come as a surprise, though 300-350W is more believable since the TSMC N4 (4nm) node is retained.
The GeForce RTX 5080 will feature the GB203 GPU which is half as big as the GB202. It will consist of around ~10,000 shaders paired with 16GB of graphics memory across a 256-bit bus. The performance deficit versus the 5090 should be similar to the 4080-4090 delta. It is expected to deliver higher frame rates than the RTX 4090, at least in ray-traced workloads.
According to @kopite7kimi, the RTX 5080 will launch ahead of the RTX 5090, likely this fall. NVIDIA tends to hold private events for such launches, so don’t expect anything at Computex. Traditionally, the 80 and 90-class cards are the first to arrive (together), followed by the 70 and 60-class SKUs.
