Nvidia announces RTX 50-series graphics cards with DLSS 4

Nvidia today announced the GeForce RTX 50 series, the company’s next generation of GPUs for desktop and mobile computing. Based on the latest Blackwell The current product stack includes the RTX 5090, RTX 5080, RTX 5070 Ti, and RTX 5070 in both desktop and mobile variants.

The four GPUs have obvious differences in specs, but let’s focus on the commonalities first. All new Blackwell GPUs are based on TSMC 4NP technology. They are equipped with the new 5th generation Tensor cores and 4th generation ray tracing cores. There is a new 9th generation NVENC encoder and 6th generation NVDEC decoder.

All four announced models feature faster GDDR7 memory. Connectivity has improved, with all cards supporting PCIe 5.0, as well as the latest DisplayPort 2.1b UHBR20 (80 Gbps) and HDMI 2.1b (48 Gbps) ports. All 50-series cards will also support DLSS 4, as well as the upcoming Reflex 2 with Frame Warp.

The flagship RTX 5090 card has 32% more CUDA cores than the RTX 4090, as well as 154% more AI TOPS and 66% more ray tracing teraflops. You also get 8GB more VRAM, which is also faster on the 512-bit memory bus. Power consumption has also increased, with Nvidia targeting 575W for the 5090, although the new Founders Edition card somehow only uses a dual-slot design with dual flow fans.

The RTX 5090 retails for an incredible $1,999, which is 33% more expensive than the RTX 4090. It will be available starting January 30th.

The RTX 5080 features a modest 5% increase in CUDA core count over the 4080 Super and similarly smaller increases in tensor and ray tracing performance over the 5090. Memory also remains at 16GB, but with faster GDDR7 on the same 256-bit memory bus. As you can imagine, power also increased relatively little, from 320W on the 4080 Super to 360W.

The RTX 5080 is priced at $999, the same as the RTX 4080 Super and less than the RTX 4080’s $1,199 starting price. The 5080 will also be available starting January 30th.

Finally, the RTX 5070 Ti retails for $749, which is $50 less than the launch price of the 4070 Ti and 4070 Ti Super, while the RTX 5070 costs $549, which is $50 less than the 4070 Super and matches the price of the 4070 5070 Ti and 5070 will be available in February.

Nvidia also announced DLSS 4, the latest generation of Nvidia’s Deep Learning Super Sampling image reconstruction technology. The main new component of DLSS 4 is multi-frame generation, which involves inserting one AI-generated frame before generating and inserting three AI-generated frames between every two traditionally generated frames. This method uses significantly improved tensor processing of 50-series cards to create new frames with higher accuracy and quality than before. Therefore, this feature is only available on 50 series cards.

However, Nvidia is also improving DLSS performance on older RTX cards. The new and improved single frame generation feature improves the quality of the single frame generation feature on RTX 50 and 40 series cards, further increasing frame rates while also reducing VRAM usage.

The company is also moving from a CNN-based model to a transformer-based model for features such as ray reconstruction, super resolution and DLAA. The result is an image with greater detail and greater temporal stability, reducing artifacts such as ghosting, aliasing, and blurring in reconstructed images. These new DLSS 4 improvements will appear on all RTX cards starting with the 20 series. Users can also simply enable these features in existing games through the Nvidia app without waiting for game updates.

DLSS 4 with multi-frame generation will be available in 75 apps and games, including Unreal Engine 5, from day one. Users will also be able to enable this feature in games that have not been updated through the Nvidia app using the new DLSS Override feature.