At CES 2025, Nvidia has officially lifted the veil on its new graphics cards (GPU) from the RTX 50 range, succeeding the RTX 40 series launched a little over two years ago. After numerous leaks and rumors, Blackwell GPUs are now confirmed with RTX 5090, RTX 5080, RTX 5070 Ti et RTX 5070.
Four models for all needs
The RTX 50 series comes in four main models, each aimed at satisfying different segments of the graphics card market. Nvidia thus announced (the prices in euros are not yet known):
- RTX 5090 at $1,999
- RTX 5080 at $999
- RTX 5070 Ti at $749
- RTX 5070 at $549
The RTX 5090 and RTX 5080 will be available from January 30, 2025, while the RTX 5070 Ti and RTX 5070 will arrive in February. These models are based on the all-new Blackwell architecture and benefit from GDDR7 memory, as well as a redesigned design for the Founders Edition version, which stands out with dual-flow fans and a 3D vapor chamber.
Power for the RTX 5090
Nvidia’s new flagship model, the RTX 5090, stands out for its ability to offer performance up to twice that of the RTX 4090. This card integrates 32 GB of GDDR7 RAM, a memory bandwidth of 1,792 GB /s and no less than 21,760 CUDA cores. Although its graphics power reaches 575 watts, it is more compact than the RTX 4090, with a two-slot format, allowing its installation in small PCs. Thanks to the optimization of energy consumption, Nvidia ensures that the RTX 5090 will be more efficient, although its power of 575 watts is higher than that of the RTX 4090 of around 125 watts.
In terms of performance, the RTX 5090 was demonstrated at CES by running Cyberpunk 2077 with DLSS 4 at 238 FPS, compared to 106 FPS for the RTX 4090 with DLSS 3.5, while activating ray tracing.
The other models are also efficient
Nvidia’s other graphics cards are not left out. The RTX 5080 is advertised to be twice as fast as the RTX 4080, with 16 GB of GDDR7 RAM, 960 GB/s bandwidth and 10,752 CUDA cores. It has a graphics power of 360 Watts and requires a power supply of 850 Watts. The RTX 5070 Ti and RTX 5070 target a wider audience, with more “affordable” prices. The RTX 5070 Ti, equipped with 16 GB of GDDR7 RAM and 8,960 CUDA cores, should offer twice the performance of the RTX 4070 Ti.
Software innovations with DLSS 4
One of the big announcements of this new range of GPUs concerns DLSS 4, which offers improvements in image rendering. This new Deep Learning Super Sampling system goes beyond improving frame rates, enabling the generation of up to three additional frames per traditional frame, which can increase performance by eight times. DLSS 4 uses transformer-based AI models to reduce ghosting, improve moving details, and optimize visual quality in real-time.
This technology also comes with RTX Neural Shaders and RTX Neural Faces, which aim to improve the quality of textures and faces in games, leveraging AI to generate highly accurate visual elements.
RTX 50 GPUs for laptops
Nvidia also announced the arrival of RTX 50 GPUs for laptops, with versions adapted to the performance and energy consumption constraints of mobile machines. The RTX 5090 for laptops, with its 24 GB of GDDR7 RAM, will debut in March 2025, followed by the RTX 5080, 5070 Ti and 5070.
Whether for gamers, content creators or high-performance computing professionals, the RTX 50 series promises to meet the highest demands while paving the way for a new era of AI and graphics rendering.
As a reminder, AMD announced its RX 9070 and RX 9070 XT graphics cards yesterday, but details are much lighter on them.
