Nvidia Expands Physical AI Capabilities with New Initiatives at CES 2025

Nvidia has taken its artificial intelligence expertise beyond data centers and offices, expanding into factories and on the road as it unveiled new AI initiatives at CES 2025. The company’s latest generation AI processor, Blackwell, is now in full production, and desktop graphics processing units based on this architecture will be available for gamers and creators this month.

CEO Jensen Huang announced Nvidia’s efforts to advance “physical AI,” which encompasses autonomous vehicles, factories, and robots, including humanoid robots. This marks a shift from the current generative AI phase to an “agentic AI” phase where software and devices work on behalf of users.

Nvidia Cosmos, a computing platform for accelerating physical AI development, was launched at CES 2025. This platform offers world foundation models to help developers create next-generation autonomous vehicles and robots. Ride-sharing giant Uber is among the first to adopt Cosmos.

The company also introduced agentic AI blueprints to automate work in enterprises. These specialized agents can reason, plan, and take action for developers. Additionally, Nvidia announced updates to its Omniverse simulation platform and new tools for RTX AI PCs.

Toyota, Aurora, and Continental have joined the list of mobility firms developing consumer and commercial vehicle fleets on Nvidia accelerated computing and AI. The company also introduced Project Digits, a high-end personal computer powered by its Nvidia GB10 Grace Blackwell Superchip and running the Linux-based Nvidia DGX operating system. This system will be marketed to AI researchers, data scientists, and students.

Nvidia’s stock has rallied ahead of the news, clearing an early entry point from a double-bottom base according to IBD chart analysis. The company is on six IBD stock lists: IBD 50, Leaderboard, SwingTrader, Big Cap 20, Sector Leaders, and Tech Leaders.

Source: https://www.investors.com/news/technology/nvidia-stock-ces-2025-news-physical-ai