In a bold move to democratize artificial intelligence, Nvidia has introduced Project DIGITS, a personal AI supercomputer designed to provide researchers, data scientists, and students with cutting-edge AI capabilities. First unveiled at CES 2025, this innovation is now nearing its market release, promising to bring unprecedented AI processing power to individual users.
AI supercomputing at your fingertips
At the heart of Project DIGITS is the Nvidia GB10 Grace Blackwell Superchip, a revolutionary system-on-chip (SoC) that combines Nvidia’s Blackwell GPU with the fifth-generation Tensor Cores and a high-efficiency Grace CPU. This integration enables up to 1 petaflop of AI performance with FP4 precision, all while operating within standard power consumption limits.
“AI is becoming mainstream across all industries and applications. With Project DIGITS, the Grace Blackwell Superchip is now in the hands of millions of developers,” said Jensen Huang, Nvidia’s CEO. “Placing AI supercomputing power on the desks of data scientists, AI researchers, and students will enable them to actively shape the AI era.”
Efficiency and scalability
Project DIGITS is not just about raw power—it’s also about efficiency and scalability. With 128GB of unified memory and up to 4TB of NVMe storage, the system is capable of running AI models with up to 200 billion parameters. For even larger workloads, two DIGITS units can be connected via Nvidia ConnectX networking, allowing them to process models with up to 405 billion parameters.
In collaboration with MediaTek, a leader in ARM-based SoC design, Nvidia has optimized DIGITS for low power consumption and high efficiency, ensuring that advanced AI tasks can be executed without the need for massive data center infrastructure.
Seamless transition from local to cloud computing
One of Project DIGITS’ standout features is its ability to bridge the gap between local AI development and large-scale deployment. Users can prototype and fine-tune models on their local workstations using the Linux-based Nvidia DGX OS and then seamlessly deploy them to Nvidia DGX Cloud, AI-optimized cloud platforms, or enterprise data center infrastructure.
Additionally, DIGITS provides access to Nvidia’s extensive AI software ecosystem, including:
- Nvidia NeMo for fine-tuning AI models,
- RAPIDS libraries for data science acceleration,
- Popular AI frameworks such as PyTorch and Jupyter Notebooks.
- Nvidia Blueprints and NIM microservices, enabling developers to build AI-driven applications with enterprise-level support.
Making AI accessible to everyone
The launch of Project DIGITS aligns with Nvidia’s broader strategy to meet the growing demands of generative AI and large language models (LLMs). The Grace Blackwell architecture, named after mathematician David H. Blackwell, represents a major leap in AI computing. With 208 billion transistors—2.5 times more than previous Nvidia Hopper GPUs—the Blackwell architecture is the largest GPU ever built, delivering 20 petaflops of processing power.
With support for next-generation Transformer Engines and advanced NVLink interconnects, DIGITS is optimized for real-time AI applications across industries, from healthcare to automotive technology.
Availability and pricing
Project DIGITS is set to launch in May 2024, with prices starting at $3,000. This combination of affordability and cutting-edge performance positions DIGITS as a game-changer in AI development, making supercomputing power accessible to individuals rather than just large enterprises.
By placing AI supercomputers directly in the hands of developers, Nvidia is not just advancing AI accessibility—it is accelerating a new wave of innovation that could reshape industries and redefine what is possible in the AI era.
