NVIDIA Launches Project DIGITS: The Personal AI Supercomputer That’s Right at Your Desk | Cpu hardware examples | Gpu hardware laptop | 4 main parts of a computer | Turtles AI

NVIDIA unveiled Project DIGITS, a personal AI supercomputer that leverages the new Grace Blackwell architecture-based GB10 superchip, delivering advanced computing power for researchers, data scientists, and developers. The device is designed to run complex models with 200 billion parameters, making AI accessible to everyone.

Key Points:

• Project DIGITS is NVIDIA’s new personal AI supercomputer.
• Featuring the Grace Blackwell GB10 superchip, it delivers 1 petaflop of AI performance.
• Flexibility in running AI models, with the ability to prototyping locally and deploy to the cloud.
• Available in May 2025, starting at $3000.

NVIDIA today launched Project DIGITS, an AI supercomputer designed to put the power of the Grace Blackwell platform in the hands of developers, researchers, and students around the world. Powered by the GB10 superchip, a SoC based on the innovative NVIDIA Grace Blackwell architecture, this device delivers petaflop performance, capable of running large AI models with up to 200 billion parameters. Featuring a powerful NVIDIA Blackwell GPU with fifth-generation CUDA cores and Tensor Cores, the GB10 is paired with a high-performance NVIDIA Grace CPU with 20 cores designed with the Arm architecture, for efficient resource management and exceptional power efficiency. This design has also been optimized through collaboration with MediaTek, a leader in Arm SoCs, to deliver high performance at low power. The combination of these components makes Project DIGITS the most compact and low-power AI supercomputer capable of being used in a typical workstation.

Project DIGITS represents a unique opportunity for AI professionals to develop, refine and test models directly on a desktop system, without the need for complex or expensive infrastructure. The system is ready to work with accelerated clouds or data centers, enabling model deployment at scale with compatibility with NVIDIA DGX Cloud and NVIDIA AI Enterprise platforms. Developers can also leverage a broad ecosystem of software tools and libraries, including machine learning frameworks like PyTorch and Python, and models available through the NVIDIA NGC catalog. Experimentation options are expanded with NVIDIA NeMo for language model refinement and NVIDIA RAPIDS libraries for faster data science. Additionally, the high-speed NVIDIA ConnectX network enables multiple Project DIGITS systems to connect to handle even more demanding workloads, running models with up to 405 billion parameters.

The device also includes 128GB of unified memory and up to 4TB of NVMe storage, a level of resources that allows developers to tackle the most advanced computational challenges. With a scalable architecture, the system allows for quick and easy prototyping, and then easily expand models and applications to cloud infrastructures.

The project is aimed at a wide audience, from data scientists and researchers to students who want to enter the world of AI. Available in May 2025 with a starting price of $3,000, Project DIGITS aims to democratize access to the computing power needed to fuel the AI ​​revolution. The product will be sold by NVIDIA and major partners, offering a scalable and powerful solution to tackle AI challenges in every industry.

With availability imminent, developers can already sign up for updates and take advantage of the opportunities offered by the world’s smallest and most powerful personal AI supercomputer.