This week at CES 2025, Nvidia unveiled a suite of products that signal its ambitions far beyond its traditional strongholds in graphics cards and data centers.

The $3,000 AI supercomputer

Leading Nvidia's announcements is Project DIGITS, a desktop AI supercomputer priced at $3,000 - comparable to a high-end gaming laptop. This device packs Nvidia's GB 10 Grace Blackwell Superchip with 128GB memory and can run AI models up to 200 billion parameters locally. When linked together, two units can handle models up to 405 billion parameters.

The pricing and positioning suggest Nvidia aims to democratize AI development, bringing powerful compute capabilities previously confined to data centers directly to developers' desks. This move also may help defend Nvidia's dominant CUDA framework against emerging competitors.

Gaming gets an AI upgrade

Nvidia hasn't forgotten its gaming roots. The company's new GeForce RTX 50 Series graphics cards feature the Blackwell Neural Rendering Architecture with 92 billion transistors. But perhaps more intriguingly, Nvidia is transforming non-player characters (NPCs) in games through its expanded ACE technology.

In PUBG, players can now team up with AI-powered teammates that make strategic decisions and communicate naturally. These aren't simple scripted bots - they use small language models and multi-modal AI to perceive the game environment and make human-like decisions in real-time.

Autonomous everything 

Transportation represents another major frontier. Nvidia announced partnerships with Toyota, Aurora, and Continental to provide AI computing technology for autonomous vehicle fleets. Toyota will integrate Nvidia's Drive AGX Orin system-on-chip, while Aurora and Continental are collaborating on driverless trucks targeted for mass production by 2027.

Supporting these efforts is the new Omniverse Cloud Sensor RTX, which uses generative AI to simulate various sensors for autonomous system development. This addresses a critical challenge in collecting diverse training data, especially for rare or hazardous scenarios.

New models

Underpinning many of these advances is Nemotron, Nvidia's new family of AI models designed for sophisticated enterprise tasks. Available in three sizes - Nano, Super, and Ultra - these models excel at instruction following, chat functionality, coding, and mathematical operations. Early adopters include SAP and ServiceNow, suggesting Nvidia's serious push into enterprise software integration.

Looking ahead 

Nvidia's announcements reflect a company leveraging its AI expertise to expand into new markets while strengthening its core businesses. The $3,000 price point for Project DIGITS could spark a new era of local AI development, while partnerships with major automakers position Nvidia to capitalize on the autonomous vehicle revolution.

We publish daily research, playbooks, and deep industry data breakdowns. Learn More Here