At CES 2025 in Las Vegas, Nvidia unveiled Project Digits, a “personal AI supercomputer” that provides access to the company’s Grace Blackwell hardware platform in a compact form factor. Designed for AI researchers, data scientists, and students, Project Digits packs Nvidia’s new GB10 Grace Blackwell Superchip, which provides up to petaflops of computing performance for prototyping, fine-tuning, and running AI models. Nvidia claims a single Project Digits unit can run models up to 200 billion parameters in size. Parameters are roughly proportional to the problem-solving skill of the model, and models with more parameters generally perform better than models with fewer parameters. The GB10, developed in partnership with MediaTek, features an Nvidia Blackwell GPU connected to a 20-core Nvidia Grace CPU. On the Project Digits enclosure, the chip is connected to a 128GB memory pool and up to 4TB of flash storage. Nvidia says that two Project Digits machines can be connected together to run up to 405 billion parameter models, if a project takes place. But they are not cheap. The Project Digits machine, which runs Nvidia’s Linux-based DGX OS, will be available starting in May from “premium partners” for $3,000, the company said. So, not just anyone can pay for their own Project Digit unit. Nvidia CEO Jensen Huang thinks there is a market, however. “With Project Digits, the Grace Blackwell Superchip is coming to millions of developers,” he said in a statement. “Putting an AI supercomputer on the desk of every data scientist, AI researcher and student empowers them to participate and shape the age of AI.”
