At the ongoing CES, Nvidia unveiled Project Digits, a personal AI supercomputer designed to handle AI models with up to 200 billion parameters.
Set to launch in May 2025, this compact powerhouse is built around the new GB10 Grace Blackwell Superchip, a marvel of engineering that delivers the performance needed to handle sophisticated AI models, all while running on a standard power outlet. Starting at just $3,000, Project Digits brings cutting-edge AI capabilities within reach of researchers, developers, and even students.
At first glance, Project Digits might remind you of a Mac Mini, but don’t let its size fool you. This system is capable of handling AI models with up to 200 billion parameters. For context, that’s leagues ahead of most consumer-grade computing hardware. With 128GB of unified memory (most laptops max out at 32GB) and 4TB of NVMe storage, Project Digits is built for serious AI development. For users with even higher demands, two systems can be linked together to handle up to 405 billion parameters, rivaling the capabilities of Meta’s latest AI model, Llama 3.1.
Built for AI, Designed for Everyone
Nvidia’s CEO Jensen Huang envisions Project Digits as a game-changer. “Placing an AI supercomputer on the desks of every data scientist, AI researcher, and student empowers them to engage and shape the age of AI,” he said during the launch. It’s a bold step toward democratizing AI development, making the tools needed for breakthroughs more accessible to innovators worldwide.
The GB10 chip at the heart of Project Digits delivers an astonishing 1 petaflop of AI performance, meaning it can perform 1 quadrillion calculations per second. Built with Nvidia’s latest CUDA and Tensor cores and featuring a power-efficient 20-core Arm-based Grace CPU, the system combines raw power with energy efficiency. MediaTek collaborated on its design, optimizing it for peak performance and usability.
Seamless AI Development
Project Digits isn’t just about hardware; Nvidia has equipped it with an ecosystem of AI tools. Developers can use popular frameworks like PyTorch and Python, refine models with the Nvidia NeMo framework, and accelerate workflows with RAPIDS libraries. The system also integrates smoothly with Nvidia’s cloud-based AI Enterprise platform, letting users develop locally and scale effortlessly.
With Project Digits, Nvidia bridges the gap between affordability and high-end AI performance, enabling developers and startups to innovate without limits. The AI revolution has never been more accessible.
