Through a presentation speech delivered by its CEO at Les 2025 in Las Vegas on Monday, NVIDIA made a number of new announcements.
The main function was the GeForce RTX 50-series, the company’s second flagship desktop GPU delivering in over two years. The design accounts have the power to run AI versions and can use DLSS 4 neurological putting technology to increase entertainment frame rates by up to eight times.
Also, Jensen Huang, wearing his name leather coat, unveiled a new NVIDIA computer for AI researchers, a methodology for human robots that allows them to learn by copy, a world foundation model platform, and a partnership with Toyota for autonomous vehicles.
GeForce RTX 50-series pc GPUs
The RTX 5090, which has 92 billion circuits and generates 3, 352 trillion AI procedures per second, is the most developed of the new GeForce line.
” The GeForce RTX 50 Set delivers breakthroughs in AI-driven processing, including neurological graphics, modern human technology, geometry and lighting”, NVIDIA claims in a press release.
The RTX 5090 will start at$ 1, 999, marking a$ 400 increase over its predecessor, the RTX 4090, but it is also twice as fast. The RTX 5080 delivers 1, 800 TOPS and will start at$ 999. Both will be accessible from Jan. 30.
The RTX 5070 Ti provides 1, 400 Top, similar to the RTX 4090, and will cost$ 749. However, the RTX 5070 presents 1, 000 Top at$ 549. Both will be obtainable in February.
Acer, ASUS, Dell, HP, Lenovo, and Razer have designed devices that include the fresh RTX 50-series. March will see the release of all cards, with the exception of the RTX 5070 tablets, which will debut in April.
The most significant computer images development since the introduction of multifunction shading 25 years ago, Huang said,” Blackwell is combining AI-driven neural rendering and light tracing.”
Private computer for AI experts
On top of the new device collection, NVIDIA introduced the world to its specific Artificial supercomputer, Project DIGITS. This is intended to be used by AI analysts, data experts, and students to” recharge Artificial innovation”, according to NVIDIA.
The GB10 Grace Blackwell superchip, which consists of a GPU and a 20-core CPU, offers up to a petaflop of computing power for the DIGITS computer. It is capable of running huge language models with up to 200 billion parameters; in general, the number of parameters is equal to efficiency capacity.
Finish with 128GB of memory, away to 4TB of display store, and NVIDIA’s DGX operating system, the servers will be available in May. It will start at$ 3, 000.
Humanoid computers have the ability to mimic humans.
In March 2024, NVIDIA unveiled GR00T — a general-purpose base Artificial model for human computers. At Les 2025, Huang announced a new mode, or template, for GR00T that allows it to understand by copy.
In other words, individuals will be able to work out movements they want the robot to accomplish, it will catch them as a modern twin, and then do them consistently. An Apple Vision Pro headset is all that is required for this.
The technology is intended to facilitate the education of large numbers of robots working in manufacturing facilities. The$ 50 trillion manufacturing and logistics industries will be radically impacted by physical AI. Everything that moves, from factories and warehouses to factories and warehouses, will be embodied by AI, according to Huang.
This new blueprint builds on the Real-Time CAE Digital Twins, which was released in November, enabling the creation of interactive simulations for engineering applications. Both leverage NVIDIA’s Omniverse platform for real-time 3D collaboration.
Huang also revealed other blueprints for creating custom AI agents for businesses using NVIDIA NIMs, including microservices with the APIs, domain-specific code, optimized inference engines, and enterprise runtime required for generative AI. Additionally, he made the Llama Nemotron family of open large language models available to help with the deployment and creation of agents. TechRepublic predicted that by the end of the year, AI agents will be more popular in 2025.
World foundation models for robotics
Huang unveiled a new “world foundation model” platform. These AI models were inspired by Cosmos, the way that humans create mental representations of the world. The first Cosmos models released are designed specifically for robotics, focusing on “physics-based simulation and synthetic data generation”.
” Cosmos WFMs are purpose-built for physical AI research and development, and can generate physics-based videos from a combination of inputs, like text, image and video, as well as robot sensor or motion data”, according to NVIDIA.
SEE: NVIDIA: Blackwell Architecture Will Accelerate AI Products in Late 2024
They will all be trained on 9, 000 trillion tokens, or pieces of video data, including 20 million hours of data from autonomous driving, robotics, human interactions, and other domains. Uber and Wayve are the only companies that have committed to using Cosmos models in use cases involving self-driving cars.
The Cosmos models will be available publicly through NVIDIA, Git Hub, and Hugging Face.
Toyota and autonomous driving partnership
Finally, Huang announced a new partnership between NVIDIA and Toyota to develop next-generation autonomous vehicles. The upcoming Toyota models will run on the DriveOS operating system and NVIDIA’s DRIVE AGX Orin supercomputer.
Following Aurora and Continental’s announcement to expand the deployment of driverless trucks using NVIDIA DRIVE, autonomous vehicle companies announced a partnership. Huang projected NVIDIA’s automotive revenue to reach$ 5 billion in his keynote.
” The autonomous vehicle revolution has arrived, and automotive will be one of the largest AI and robotics industries”, he said in a press release.
