In a recent press conference at the ATCES 2025 in Las Vegas, Nvidia unveiled Project Digits, a compact form factor personal AI supercomputer that provides access to the company’s powerful Grace Blackwell hardware platform. The system is designed specifically for AI researchers, data scientists, and students who need high-performance computing capabilities for prototyping, fine-tuning, and running AI models.
A Cloud Computing Platform on Your Desk
According to Nvidia CEO Jensen Huang, Project Digits runs the entire Nvidia AI stack, including all of Nvidia’s software, which means that users can access a cloud computing platform from the comfort of their own desk. This is an exciting development for researchers and developers who need high-performance computing capabilities but may not have access to expensive data centers or large-scale computing infrastructure.
The Powerhouse Behind Project Digits
At the heart of Project Digits lies Nvidia’s new GB10 Grace Blackwell Superchip, which delivers up to a petaflop of computing performance. The Superchip features an Nvidia Blackwell GPU connected to a 20-core Nvidia Grace CPU and is accompanied by a 128GB pool of memory and up to 4TB of flash storage.
What Makes Project Digits So Special?
One of the key benefits of Project Digits is its ability to run large-scale AI models. According to Nvidia, a single unit can handle models with up to 200 billion parameters in size. Parameters are essentially a measure of a model’s problem-solving skills, and larger models generally perform better than smaller ones.
To give you an idea of just how powerful Project Digits is, consider the following:
- A 200-billion-parameter model is significantly larger than most existing AI models.
- Two Project Digits machines can be linked together to run even larger models with up to 405 billion parameters in size.
This means that researchers and developers who need high-performance computing capabilities for running large-scale AI models now have a powerful tool at their disposal.
The Cost of High-Performance Computing
While Project Digits is an exciting development, it’s worth noting that the system comes with a hefty price tag. Nvidia plans to make Project Digits available starting in May from "top partners" for $3,000 per unit. This may be out of reach for many researchers and developers who need high-performance computing capabilities.
However, according to Jensen Huang, there is a market for this kind of technology:
"With Project Digits, the Grace Blackwell Superchip comes to millions of developers. 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."
How Does Project Digits Work?
According to Nvidia, Project Digits can be used in a standalone configuration or connected to a primary Windows or Mac PC. The system runs Nvidia’s Linux-based DGX OS and is designed to provide users with access to high-performance computing capabilities from the comfort of their own desk.
Here are some key features of Project Digits:
- Runs on Nvidia’s Linux-based DGX OS
- Includes a 128GB pool of memory and up to 4TB of flash storage
- Can be linked together for even greater performance
- Provides access to Nvidia’s powerful AI software stack
Conclusion
Nvidia’s Project Digits is an exciting development in the field of AI research and development. By providing researchers and developers with high-performance computing capabilities on their own desk, Project Digits has the potential to revolutionize the way we approach AI research and development.
While the cost may be out of reach for many users, Nvidia’s commitment to making this technology available to millions of developers is a testament to the company’s dedication to advancing the field of AI.
As we move forward into an age of increasingly complex AI models, Project Digits has the potential to play a major role in shaping the future of AI research and development.
