NVIDIA's Project Digits is a $3,000 personal AI supercomputer
What's the story
Leading chipmaker NVIDIA has announced the launch of a personal AI supercomputer called Project Digits. The announcement was made at CES and the product will be available in May.
At the heart of Project Digits lies the GB10 Grace Blackwell Superchip, a compact yet powerful processor that can run complex AI models from a standard power outlet.
It is designed to empower researchers, data scientists, and students, providing high-performance computing capabilities to individual users.
Product details
Project Digits: A desktop-sized AI powerhouse
Project Digits, which comes at a price of $3,000 (roughly ₹2.3 lakh), is a desktop-sized system capable of processing AI models with as many as 200 billion parameters.
Until now, this kind of processing power was only possible with larger and more energy-consuming systems.
The product's design is also reminiscent of Mac Mini, making it a compact addition to any workspace.
Vision
NVIDIA CEO envisions AI supercomputers for all
NVIDIA's CEO, Jensen Huang, envisions a future where AI is mainstream in every app for every industry.
He believes Project Digits will bring the Grace Blackwell Superchip 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," Huang said in a press release.
Tech specs
Project Digits: Advanced specifications and capabilities
Each Project Digits system packs 128GB of unified, coherent memory and up to 4TB of NVMe storage.
For more intensive applications, two systems can be connected to run models with up to 405 billion parameters.
The GB10 chip offers up to 1 petaflop of AI performance at FP4 precision, with NVIDIA's latest-generation CUDA cores and fifth-generation Tensor Cores linked via NVLink-C2C to a Grace CPU with 20 power-efficient Arm-based cores.
Collaboration
MediaTek collaborates with NVIDIA on GB10 chip development
MediaTek, known for its Arm-based chip designs, worked with NVIDIA to develop the GB10 chip. The goal was to optimize power efficiency and performance.
Users of Project Digits will also get access to NVIDIA's AI software library, including development kits, orchestration tools, and pre-trained models available through the NVIDIA NGC catalog.
Software support
Project Digits supports popular AI development frameworks
The system runs on Linux-based NVIDIA DGX OS and supports popular frameworks such as PyTorch, Python, and Jupyter notebooks.
Developers can fine-tune models using the NVIDIA NeMo framework and accelerate data science workflows with NVIDIA RAPIDS libraries.
Users can develop and test their AI models locally on Project Digits, then deploy them to cloud services or data center infrastructure with the same Grace Blackwell architecture and NVIDIA AI Enterprise software platform.