NVIDIA unveils Project DIGITS: A $3,000 personal AI supercomputer

At CES 2025, NVIDIA unveiled Project DIGITS, a personal AI supercomputer designed to make cutting-edge AI computing accessible to researchers, data scientists, and students worldwide. Powered by the revolutionary NVIDIA GB10 Grace Blackwell Superchip, the platform delivers a petaflop of AI performance, enabling the development, fine-tuning, and deployment of large-scale AI models directly from a desktop.

“AI will be mainstream in every application for every industry. With Project DIGITS, the Grace Blackwell Superchip comes to millions of developers,” said Jensen Huang, NVIDIA founder and CEO. “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.”

At the heart of Project DIGITS is the GB10 Grace Blackwell Superchip, a system-on-a-chip (SoC) that delivers up to 1 petaflop of AI performance at FP4 precision.

Key features of the GB10 include:

    •    Advanced Architecture: Built on NVIDIA’s Grace Blackwell platform, it combines a high-performance NVIDIA Blackwell GPU with fifth-generation Tensor Cores and an NVIDIA Grace CPU featuring 20 power-efficient cores based on Arm architecture.
    •    Exceptional Power Efficiency: Developed in collaboration with MediaTek, the superchip offers industry-leading efficiency, performance, and connectivity while operating on standard electrical outlets.
    •    Unified Memory: 128GB of memory and up to 4TB of NVMe storage enable users to process large datasets and run AI models with up to 200 billion parameters.
    •    Scalability: Two DIGITS systems can connect via NVIDIA ConnectX® networking to handle up to 405 billion parameters.

Project DIGITS allows developers to prototype AI models locally on Linux-based systems running NVIDIA DGX OS and scale seamlessly to cloud or data centre environments via NVIDIA DGX Cloud™. This streamlined workflow uses the same Grace Blackwell architecture and NVIDIA AI Enterprise software platform for consistency across development stages.

Users can access NVIDIA’s extensive AI software library, including tools like:

    •    NVIDIA NeMo: For fine-tuning models.
    •    NVIDIA RAPIDS: To accelerate data science tasks.
    •    Frameworks like PyTorch and platforms like Jupyter notebooks for a comprehensive development experience.

For developers focused on agentic AI, NVIDIA Blueprints and NVIDIA NIM™ microservices are available for research and testing via the NVIDIA Developer Program.

Starting at $3,000, Project DIGITS brings AI supercomputing within reach for a broader audience. It will be available from NVIDIA and its partners beginning in May 2025, with sign-ups for notifications now open.