At CES 2025, Nvidia not only announced its upcoming RTX 50 series but also introduced the first-ever personal AI supercomputer, Project Digits. With a starting price of $3,000, Nvidia is targeting Digits towards developers and data scientists seeking an affordable and accessible hardware and software platform to build end-to-end generative AI solutions. This accessibility marks a significant step in making AI training more available to individuals and smaller organizations.

Nvidia has already played a major role in the AI boom, and with Digits, the idea of bringing the power of an AI supercomputer to the desktop addresses the growing need for enhanced AI performance in a compact and accessible form factor. With Digits, startups can tackle complex AI challenges without relying on expensive cloud computing resources or large-scale supercomputing infrastructure. From new AI applications to fine-tuning LLMs for specific tasks to researching new AI algorithms and architectures, Nvidia’s Project Digits is poised to bring a groundbreaking revolution to the AI race.

A new SoC based on the NVIDIA Grace Blackwell architecture

At the heart of this AI supercomputer lies the Nvidia GB10 Grace Blackwell Superchip, which combines a powerful Blackwell GPU with a 20-core Grace CPU. This engineering marvel is what enables Digits to handle complex AI tasks quickly and efficiently. Additionally, Digits boasts 128GB of unified memory and high-speed NVMe storage, making it particularly well-suited for AI workloads that often involve large datasets and complex computations.

Key specifications:

• Blackwell GPU: Includes CUDA cores and fifth-generation Tensor Cores for accelerating AI calculations.
• Nvidia Grace CPU: Features 20 power-efficient Arm cores complementing the GPU for balanced AI workloads.
• NVLink-C2C interconnect: Provides a high-bandwidth, low-latency connection between the GPU and CPU for efficient data transfer.
• 128GB of unified memory: Shared memory pool for CPU and GPU, eliminating data copying and speeding up processing.
• High-speed NVMe storage: Ensures rapid access to data for AI model training and execution.
• 1 petaflop of AI performance: Handles complex AI tasks and large AI models.
• Power-efficient design: Operates using a standard wall outlet.

With such a powerful set of specifications packed inside, you might think that Digits would require tons of power. However, unlike traditional supercomputers that need specialized power and cooling infrastructure, Digits is designed with power efficiency in mind. In fact, it can operate using a standard wall outlet, eliminating the need for additional power sources. This impressive power efficiency was achieved thanks to Nvidia’s collaboration with MediaTek, which helped design an energy-efficient CPU specifically for this AI supercomputer.

Project Digits also comes preloaded with Nvidia's comprehensive AI software stack, such as the Nvidia AI Enterprise software platform, along with a wide array of familiar tools and frameworks essential for AI research and development. While Digits runs on a Linux-based operating system, it is designed to seamlessly integrate with Windows PCs through Windows Subsystem for Linux technology. This allows developers working primarily in Windows environments to easily leverage the power of Digits for their AI projects.

Digits aims to make AI accessible to everyone, everywhere

With its exceptional AI performance, integrated software stack, power efficiency, and, most importantly, a small form factor, Digits is capable of handling massive AI models with 200 billion parameters. For context, this was previously only achievable by large-scale supercomputers. The fact that Digits can bring this level of performance to desktops shows just how revolutionary it will be for enabling developers to experiment with and deploy cutting-edge AI models locally. Nvidia Digits is expected to be available in May from Nvidia and its partners, with a starting price of $3,000. At such a competitive price, Digits is sure to attract a lot of attention and will undoubtedly become a viable option for a broader range of users.

Source: Nvidia

About the Author

Hassam Nasir

Hassam boasts over seven years of professional experience as a dedicated PC hardware reviewer and writer.