For the first time, Nvidia is promoting its new GeForce RTX 40 SUPER GPUs with AI performance in generative AI applications.
At this year’s CES in Las Vegas, Nvidia announced several new products and partnerships in gaming, content creation, and robotics. These include new GPUs for gamers and content creators, generative AI models for more realistic characters in games and applications, and advances in robotics.
Of particular interest to gamers and AI hobbyists, the company unveiled the new GeForce RTX 40 SUPER GPU series, which is said to offer more performance than its predecessors. Nvidia’s announcement follows the trend of the last few years and focuses on the numerous AI features supported by the hardware. For the first time, there are also concrete figures for AI performance: the RTX 4080 SUPER is expected to achieve 836 AI TOPS, making it significantly faster in generative AI applications such as Stable Diffusion in addition to DLSS with Frame Generation. In AI workloads, the GeForce RTX 4080 SUPER generates video over 1.5x faster and images over 1.7x faster than the RTX 3080 Ti. How this performance compares to the rest of the RTX 40-series remains to be seen.
RTX 4080 SUPER should be 1.4x faster than RTX 3080 Ti
According to Jeff Fisher, senior vice president of GeForce at Nvidia, the GeForce RTX 4080 SUPER supports fully ray-tracing games in 4K and is 1.4 times faster than the RTX 3080 Ti in graphics-intensive games, which is about 20 percent faster than the regular 4080. The RTX 4080 SUPER will be available starting January 31 for $999.
The RTX 4070 Ti SUPER was also introduced. Compared to its predecessor, the company has added more cores, increased the frame buffer to 16 GB, and expanded the memory bus to 256 bits. The RTX 4070 Ti SUPER is said to be 1.6 times faster than a 3070 Ti. The RTX 4070 Ti SUPER will be available starting January 24th for 799 US dollars.
The RTX 4070 SUPER was also announced, which should be able to beat an RTX 3090. It will be available from January 17 for 599 US dollars.
Nvidia also announced a “new wave” of RTX laptops that will be available from all major manufacturers.
Generative AI for NPCs and a GenAI model from Getty Images
The company also unveiled an update to the Avatar Cloud Engine (ACE), a platform that brings digital avatars to life with generative AI. ACE AI models can be run in the cloud or locally on a PC. New ACE Production Microservices for Audio2Face and Riva Automatic Speech Recognition are now available for developers to plug into their pipelines. Developers using ACE include Convai, Charisma.AI, Inworld, miHoYo, NetEase Games, Ourpalm, Tencent, Ubisoft, and UneeQ.
In addition, the RTX Remix application will be released in open beta on January 22. RTX Remix allows modders to rework classic games with full ray tracing, DLSS, Nvidia Reflex, and generative AI texture tools that convert low-resolution textures into physically correct 4K materials.
As part of the presentation, Fisher also announced that iStock by Getty Images is launching its alternative to DALL-E 3, Midjourney, and Firefly. The iStock service allows users to create 4K images from text based on Getty Images’ extensive catalog of licensed, copyrighted creative content. New editing APIs will also be available soon, giving customers powerful control over the generated images. The generative AI service is available now on istock.com, and advanced editing capabilities will be released via APIs.
Nvidia also announced new partnerships in the automotive industry and provided insight into the role generative AI will play in robotics.