What is Project DIGITS
Project DIGITS is a personal AI supercomputer launched by NVIDIA, providing powerful computing capabilities for AI researchers, data scientists, and students. Project DIGITS is based on the NVIDIA Grace Blackwell architecture and is equipped with the GB10 Superchip, delivering up to 1 quadrillion AI operations per second and supporting large language models with up to 200 billion parameters. Project DIGITS features 128GB of unified memory and up to 4TB of NVMe storage, allowing users to develop and run AI models locally, with support for seamless deployment to the cloud or data centers. Project DIGITS supports a variety of NVIDIA AI software tools, helping developers prototype, fine-tune, and test models, driving AI innovation. Project DIGITS will be available in May, starting at $3,000.
Main Features of Project DIGITS
- Powerful AI Computing Capability: Equipped with the GB10 Superchip, delivering up to 1 quadrillion AI operations per second and supporting large language models with up to 200 billion parameters.
- Local Development and Deployment: Users can develop and run AI models on their local desktop systems, with support for seamlessly deploying models to cloud or data center infrastructure.
- Rich Software Support: Provides a broad range of NVIDIA AI software libraries, including software development kits, orchestration tools, frameworks, and models, supporting the development and testing of various AI applications.
- Efficient Memory and Storage: Features 128GB of unified memory and up to 4TB of NVMe storage, ensuring high efficiency in data processing and model operation.
- Flexible Connectivity: Based on NVIDIA ConnectX® networking, two Project DIGITS units can be connected together to further enhance computing power and support larger-scale model operation.
Technical Specifications of Project DIGITS
- GB10 Superchip：
- Architecture: Based on the NVIDIA Grace Blackwell architecture, integrating a high-performance NVIDIA Blackwell GPU and NVIDIA Grace™ CPU.
- Performance: Delivers up to 1 quadrillion AI computing operations per second, supporting FP4 precision computation.
- Connectivity: Uses NVLink®-C2C chip-to-chip interconnect technology to achieve efficient communication between the GPU and CPU.
- Power Efficiency: Based on advanced design and manufacturing processes, achieving a balance between high performance and low power consumption.
- Unified Memory Architecture: Equipped with 128GB of unified, coherent memory, supporting efficient memory access and data sharing, reducing data transfer latency, and improving computing efficiency.
- Storage and Scalability：
- Provides up to 4TB of NVMe storage, supporting fast data read/write and storage.
- Based on NVIDIA ConnectX® network connectivity, enabling multiple devices to work together, supporting larger-scale model operation and computing tasks.
Project DIGITS Official Website Address
- Project Official Website：https://www.nvidia.com/en-eu/project-digits/
Application Scenarios of Project DIGITS
- AI Research and Development: Researchers use the powerful computing capability to develop and experiment with deep learning models, exploring new AI algorithms and applications.
- Data Science and Analysis: Data scientists conduct large-scale data analysis and train machine learning models, accelerating data processing and model optimization.
- Education and Teaching: Used as a teaching tool in universities and research institutions to help students learn AI and machine learning technologies, carry out practical projects and research.
- Corporate Innovation and Development: Enterprises carry out product development and innovation, such as developing intelligent recommendation systems, natural language processing applications, etc.
- Medical and Health: Used in the medical field for medical image analysis, disease prediction, and the development of diagnostic models, improving the accuracy and efficiency of medical diagnosis.