Nvidia GeForce RTX 50: doubles the performance and increases the autonomy for laptop PCs

Laptop PCs are also entitled to GeForce RTX 5090, 5080, 5070 Ti and 5070. Obviously, if the name is the same as the versions for desktop PCs, the characteristics of the GPUs have been changed downwards to enter the chassis of a laptop PC. So it’s hard to see how a GeForce RTX 5090 that consumes 575W can find its place in a chassis that can dissipate a maximum of 250W, not to mention that it also has a processor to be cooled.

Same names, different characteristics

The GeForce RTX 5090 for laptops includes a Blackwell GPU but has half as many CUDA units and 24 GB of GDDR7. Thus it is close to the desktop RTX 5080, but with 8 GB more video memory while using the same 256-bit bus.
Its thermal envelope can be between 80 and 150W depending on the settings of the laptop manufacturers. As with the previous generation, Nvidia allows a 25W increase depending on the use of the CPU with Dynamic Boost technology.

The mobile GeForce RTX 5080 has 7680 CUDA units of 16 GB of GDDR7 on a 256-bit bus and TGP variable from 80 to 150W depending on the integration. Unlike the desktop GeForce RTX 50 series, the mobile RTX 5080 does not see its specifications halved compared to the higher model.

The mobile GeForce RTX 5070 Ti has 5888 CUDA units with 12 GB of GDDR7 on a 192-bit bus and a power envelope of between 60 and 115W with 25W of Dynamic Boost that we found 140W that was in the previous generation.

Finally, the GeForce RTX 5070 mobile has “only” 4608 CUDA units (like the GeForce RTX 4070 mobile) 8 GB of GDDR7 on a 128-bit bus and a power envelope of between 50 and 100W.

For this generation of the mobile RTX 50 series, Nvidia has closed the gap with the previous generation GeForce RTX 5070 Ti for better performance leveling between models. The 40 series generation of laptops has a real gap between the 70 and 80 versions, which is no longer the case.

Regarding the performance, Nvidia announced the doubling of video game performance thanks to the use of DLSS 4 which introduces the generation of multiple images. The GeForce RTX 50 for laptops benefits from the same technologies as their counterparts for desktop PCs with DLSS (DLAA, Ray reconstruction, Super Resolution) and the new Transformer model as well as Reflex 2 .

Battery life up to 40% on Blackwell Max-Q

To adapt to the constraints of laptop PCs and in particular autonomy, Nvidia has implemented new energy saving mechanisms within its GPU. All its improvements bear the sweet name of Blackwell Max-Q.

Thus we find an intelligent power-off of the GPU that deactivates parts that are not needed, a fast sleep that allows the GPU to sleep constantly, even during tasks, a dynamic adaptation of frequencies that adjust frequencies depending on the task to be performed, and finally the use of more efficient GDDR7 memory.
Combined, all these optimizations allow Nvidia to increase the autonomy of laptops with Blackwell GPUs by up to 40%. To give you an idea, a Zephyrus G14 with a GeForce RTX 4070 lasted a little over 9h30 in video playback, so we can see it approaching 15 hours of battery life with a GeForce RTX 5070.

All laptop manufacturers will integrate the GeForce RTX 50 series in their models such as Asus, Acer, Dell, HP and Razer. The first expected models will be the GeForce RTX 5080 and RTX 5090 versions in March, while the RTX 5070 Ti and RTX 5070 versions will be available in April.