NVIDIA is revolutionizing the accessibility of AI development with Project DIGITS, a personal AI supercomputer that packs cutting-edge power into a desktop-friendly form factor. Featuring the new NVIDIA GB10 Grace Blackwell Superchip, Project DIGITS delivers a staggering petaflop of AI computing performance, enabling developers, researchers, and students to innovate with AI models directly from their desks. Announced at CES, this groundbreaking innovation puts the transformative potential of the Grace Blackwell architecture within reach of millions worldwide.
Unveiling the GB10 Superchip: The Heart of Project DIGITS
The NVIDIA GB10 Grace Blackwell Superchip is the core engine behind Project DIGITS. This system-on-a-chip (SoC) combines:
- NVIDIA Blackwell GPU with advanced CUDA® cores and Tensor Cores for superior AI processing.
- NVIDIA Grace CPU, an Arm-based, 20-core powerhouse designed for maximum efficiency and speed.
- NVLink-C2C connectivity, enabling seamless integration between the GPU and CPU for unparalleled performance.
With up to 1 petaflop of AI power at FP4 precision, the GB10 Superchip supports large-scale AI tasks, including running up to 200-billion-parameter language models. Notably, its energy-efficient design allows the system to run using a standard electrical outlet, democratizing access to AI innovation.
A Desktop AI Supercomputer for Everyone
Project DIGITS enables users to prototype, fine-tune, and deploy AI models with unmatched ease. Key features include:
- 128GB of unified memory for handling complex data workflows.
- Up to 4TB of NVMe storage for massive datasets and AI models.
- Expandable performance, with two systems capable of linking to support 405-billion-parameter models.
These capabilities are complemented by an intuitive development environment powered by the Linux-based NVIDIA DGX OS. Users can prototype AI locally and seamlessly scale their projects on NVIDIA DGX Cloud or other data center infrastructures, maintaining compatibility through the Grace Blackwell architecture.
Unleashing AI Development with NVIDIA’s Ecosystem
NVIDIA’s rich ecosystem further enhances Project DIGITS, offering:
- Access to AI software, frameworks, and tools via the NVIDIA NGC catalog and NVIDIA Developer portal.
- Pre-built frameworks like NVIDIA NeMo for large language models and NVIDIA RAPIDS for accelerated data science.
- Support for popular tools like PyTorch, Python, and Jupyter notebooks for streamlined workflows.
For advanced applications, developers can leverage NVIDIA Blueprints and NVIDIA NIM microservices to experiment, prototype, and eventually deploy AI solutions in production-ready environments with enterprise-grade security.
Empowering the AI Revolution
As NVIDIA CEO Jensen Huang aptly stated, “AI will be mainstream in every application for every industry.” With Project DIGITS, the Grace Blackwell Superchip places world-class AI capabilities in the hands of developers, students, and researchers everywhere, enabling them to shape the future of AI innovation.
What’s Next for AI Developers?
Whether you’re a seasoned AI researcher or a student eager to dive into AI, Project DIGITS unlocks unparalleled opportunities to explore, experiment, and scale AI solutions.
Explore NVIDIA’s Ecosystem
Ready to take your AI skills to the next level? Visit NVIDIA’s Developer Portal for resources and tools tailored to your needs.
Learn AI with Nexttrain.io
Check out Nexttrain.io’s courses to master AI development and stay ahead in the rapidly evolving tech landscape. Dive into our blog for insights on AI trends and innovations.