NVIDIA has officially launched Project DIGITS, a groundbreaking personal AI supercomputer priced at $3,000.
Aimed at AI researchers, data scientists, and students globally, Project DIGITS leverages the cutting-edge NVIDIA Grace Blackwell platform to bring unparalleled computing power to users’ desktops.
Unmatched Performance with the GB10 Superchip
At the heart of Project DIGITS is the new NVIDIA GB10 Grace Blackwell Superchip, engineered to deliver an astounding 1 petaflop of AI computing performance at FP4 precision.
| Specification | Details |
|---|---|
| Product Name | NVIDIA Project DIGITS |
| Price | $3,000 |
| Superchip | NVIDIA GB10 Grace Blackwell Superchip |
| AI Performance | Up to 1 petaflop at FP4 precision |
| GPU | NVIDIA Blackwell GPU |
| CUDA Cores | Latest-generation CUDA cores |
| Tensor Cores | Fifth-generation Tensor Cores |
| CPU | NVIDIA Grace CPU with 20 power-efficient cores |
| Memory | 128GB of unified, coherent memory |
| Storage | Up to 4TB NVMe storage |
| Model Capacity | Can run 200-billion-parameter large language models |
| Networking | NVIDIA ConnectX for linking supercomputers |
| Increased Model Capacity | Links with another Project DIGITS for 405-billion-parameter models |
| Operating System | Linux-based NVIDIA DGX OS |
| AI Software Access | Extensive library including SDKs, orchestration tools, and frameworks (NVIDIA NGC catalog) |
| Cloud Compatibility | Seamless deployment on NVIDIA DGX Cloud |
| Availability Date | May 2025 |
This system-on-a-chip (SoC) integrates a high-performance NVIDIA Blackwell GPU, which features the latest CUDA cores and fifth-generation Tensor Cores, connected through the NVLink-C2C chip-to-chip interconnect.
This innovative architecture is complemented by a powerful Grace CPU with 20 energy-efficient cores, designed in collaboration with MediaTek, a leader in Arm-based SoC designs.
Jensen Huang, founder and CEO of NVIDIA, emphasized the significance of this technology, stating, “AI will be mainstream in every application for every industry. With Project DIGITS, the Grace Blackwell Superchip comes to millions of developers.”
The project aims to empower data scientists, AI researchers, and students by placing a supercomputer directly on their desks, enabling them to prototype, fine-tune, and deploy large AI models with unprecedented ease.
Scalable AI Solutions at Your Fingertips
With Project DIGITS, users can execute inference on their models locally and then transition seamlessly to the accelerated cloud or data center infrastructure.
Each unit is equipped with 128GB of unified memory and up to 4TB of NVMe storage, allowing developers to run extensive 200-billion-parameter large language models.
Moreover, utilizing NVIDIA ConnectX networking technology, two Project DIGITS units can be linked to tackle even larger models of up to 405 billion parameters.
Project DIGITS is built on the Grace Blackwell architecture, facilitating development on local systems running the Linux-based NVIDIA DGX OS.
Once models are tested, they can seamlessly scale using NVIDIA DGX Cloud or other data center infrastructure, maintaining compatibility with the same architecture and software platform.
Developers will have access to an extensive array of NVIDIA AI software, including SDKs, orchestration tools, and frameworks from the NVIDIA NGC catalog.
Notable tools such as the NVIDIA NeMo framework for model fine-tuning and NVIDIA RAPIDS libraries for data science acceleration will be available. Additionally, users can explore NVIDIA Blueprints and microservices for AI application development.
Project DIGITS is set to launch in May 2025, with pre-orders starting at $3,000. Aspirant developers and researchers can sign up for notifications and be among the first to experience the future of AI computing.
As NVIDIA propels us into the next era of innovation, Project DIGITS stands poised to democratize access to advanced AI technologies.
ANY.RUN Threat Intelligence Lookup - Extract Millions of IOC's for Interactive Malware Analysis: Try for Free