Nvidia is poised to extend its impressive growth trajectory from 2024 into 2025, driven by a series of significant announcements made at the Consumer Electronics Show (CES) in Las Vegas on Monday. CEO Jensen Huang presented a comprehensive vision encompassing AI-powered software for robots and self-driving vehicles, alongside the unveiling of a new AI supercomputer designed for desktop use. Nvidia’s stock price surged as high as 4.7% ahead of the keynote, reflecting Wall Street’s anticipation of the company’s latest offerings. The stock has experienced a remarkable 205% increase over the past twelve months, a testament to Nvidia’s strategic investments in AI hardware, particularly its CUDA software, which empowers developers to utilize its chips for AI programming.
Nvidia’s Expanding AI Ecosystem
The core of Nvidia’s announcements centered on leveraging existing hardware, specifically its Hopper and Blackwell platforms. The company is focusing on facilitating programmer access and utilization. A key element of this strategy involves the upcoming debut of its next-generation chip, slated for presentation at the GTC conference in March. This demonstrates Nvidia’s commitment to continuously advancing its AI technology.
The GB10 Superchip: Desktop AI Power
During Monday’s event, Huang showcased Nvidia’s newest Blackwell-based chip, dubbed the GB10 superchip. This compact version, a pint-sized adaptation of the GB200 superchip, combines a Grace central processing unit (CPU) with two Blackwell graphics processing units (GPUs). The smaller GB10 pairs a Grace CPU and Blackwell GPU. Nvidia intends to make the chip available through a desktop system called Project DIGITS, equipping it with 128GB of memory and 4TB of storage. The setup is intended for research applications, facilitating “prototyping, fine-tuning, and running large AI models.” Project DIGITS will be available for $3,000 starting in May, through Nvidia and its original equipment manufacturer (OEM) partners.
Cosmos Platform: Simulating the Real World
Beyond the chip and desktop system, Nvidia introduced its open-source Cosmos platform for developing physical AI systems. This platform utilizes “world foundation models” (WFMs), which simulate real-world conditions. These physical AI systems—including humanoid robots and self-driving cars—represent a significant step towards building more versatile and adaptable AI solutions. The Cosmos platform allows companies to develop the necessary software to power these systems by simulating various usage scenarios in a virtual setting, removing the need for expensive hardware or on-road testing.
Isaac GROOT Blueprint: Accelerating Robot Development
Nvidia also debuted its Isaac GROOT Blueprint for training humanoid robots. This software, integrated with Apple’s Vision Pro headset, allows developers to perform and record specific movements for a robot to learn. Isaac GROOT Blueprint then synthesizes these movements, providing the robot with a comprehensive set of actions based on the original motions. This dramatically reduces the time needed to create future humanoid robotics systems.
Automotive Partnerships and DRIVE AGX Orin
On the automotive front, Nvidia announced a collaboration with Toyota, which will see the automaker utilize the company’s DRIVE AGX Orin chip and the Nvidia DriveOS operating system to power advanced driver assistance features in its next-generation vehicles. “The AV revolution has arrived,” Huang stated during the keynote, predicting that this would become a multi-trillion-dollar robotics industry. Nvidia has also formed agreements with Continental and self-driving truck company Aurora, who will use Nvidia’s DRIVE hardware and DriveOS software alongside Aurora’s level 4 autonomous driving system, Aurora Driver. Continental and Aurora plan to bring autonomous trucks hauling freight to roads beginning in 2027.
AI Agents and Expanded Ecosystem
Finally, Nvidia unveiled a range of AI software and hardware offerings, including its AI Blueprints, agentic AI apps. These AI agents are designed to automate tasks across different applications, mirroring the significant investments Google and Microsoft are making in this area. Nvidia additionally announced that users running its latest RTX graphics cards can now utilize foundation models via its Nvidia NIM platform. This enables AI capabilities on standard graphics cards, expanding opportunities for software developers and broadening Nvidia’s customer base.