NVIDIA has unveiled Project DIGITS, a breakthrough that puts the power of an AI supercomputer in a package small enough to fit on your desk. The new GB10 Grace Blackwell Superchip at its heart can process AI tasks as powerfully as systems that once filled entire rooms.
Think of it as having a mini research lab that plugs into a regular power outlet. The system can handle complex AI tasks that usually need massive computing centers, but now fit in a space smaller than a laptop.
“AI will be mainstream in every application for every industry,” said Jensen Huang, NVIDIA’s 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.”
The Technical Marvel Made Simple
The GB10 Superchip combines two essential parts: a 20-core Grace CPU (the brain) and a Blackwell GPU (the muscle), working together through a high-speed connection called NVLink-C2C. NVIDIA partnered with MediaTek, known for making efficient mobile phone chips, to create this powerful yet energy-efficient system.
What makes this special is its ability to deliver one petaflop of AI performance at FP4 precision – enough computing power to run AI models with up to 200 billion parameters. Two connected Project DIGITS systems can even run models up to 405 billion parameters.
Making AI Development Available to Everyone
The system runs on Linux-based NVIDIA DGX OS and comes with comprehensive AI software tools. It includes NVIDIA’s AI software stack, supporting common frameworks like PyTorch, Python, and Jupyter notebooks.
The system includes:
- Performance of one petaflop at FP4 precision
- 128GB of unified, coherent memory
- Up to 4TB of NVMe storage
- Ability to connect two systems via NVIDIA ConnectX networking
Price and Availability
Project DIGITS will be available in May, starting at $3,000. This makes professional AI development tools accessible to more developers, researchers, and students.
Real World Applications
The system allows users to develop and run inference on models using their own desktop system, then seamlessly deploy them on NVIDIA DGX Cloud, accelerated cloud instances, or data center infrastructure. Users can access NVIDIA AI software for experimentation and prototyping, including software development kits, orchestration tools, frameworks, and models from the NVIDIA NGC catalog.
What Experts Are Saying
“Our partnership with Nvidia brings together the best of both worlds—MediaTek’s efficient CPU technology and Nvidia’s powerful GPU computing. Together, we are making AI innovation more widely accessible and practical,” says Rick Tsai, MediaTek’s CEO.
Similar Posts
Raquel Urtasun, a University of Toronto computer science professor and founder of Waabi, notes that “Project DIGITS enables researchers to experiment, fine-tune, and scale solutions faster than ever.”
Looking to the Future
Project DIGITS represents a significant shift in AI development accessibility. The system supports NVIDIA’s full AI software stack and development tools, making it useful for both experienced developers and newcomers.
This development could accelerate innovation across industries by making AI development tools more widely available. Just as personal computers transformed how we work, putting AI development capabilities on more desks opens up new possibilities for innovation.
