At the CES tech trade show in Las Vegas, Nvidia CEO Jensen Huang delivered a keynote speech unveiling Project DIGITS, a revolutionary personal AI supercomputer. Powered by the cutting-edge Nvidia Grace Blackwell platform, this innovation delivers unparalleled computational power directly to AI researchers, data scientists, and students worldwide, democratizing access to advanced AI capabilities. This new system, ‘only’ 1,000 times more powerful than a typical laptop, is designed to democratize access to high-performance AI computing, enabling users to prototype, fine-tune, and deploy sophisticated AI models.
what??
NVIDIA just dropped Project DIGITS, a $3,000 personal AI supercomputer that’s small enough to look like a Mac Mini but packs 1,000x the power of your average laptop.
Handles AI models with up to 200 BILLION parameters.
This is incredible.. pic.twitter.com/z4JOeFD2JI
— el.cine (@EHuanglu) January 7, 2025
At the heart of Project DIGITS lies the Nvidia GB10 Grace Blackwell Superchip, which offers a petaflop of AI performance at FP4 precision. This superchip integrates the latest Nvidia Blackwell GPU, equipped with new CUDA cores and fifth-generation Tensor Cores, connected via the NVLink-C2C chip-to-chip interconnect to an Nvidia Grace CPU. This CPU, with 20 power-efficient cores based on Arm architecture, was co-designed with MediaTek, enhancing the superchip’s efficiency, performance, and connectivity.
The design of Project DIGITS allows it to operate from a standard electrical outlet while delivering robust performance. Each unit comes with 128GB of unified, coherent memory and up to 4TB of NVMe storage, enabling the handling of up to 200-billion-parameter large language models. Furthermore, through Nvidia ConnectX networking, two Project DIGITS systems can be combined to support models with up to 405 billion parameters, pushing the boundaries of what personal AI computing can achieve.
Project DIGITS not only offers hardware prowess but also integrates seamlessly with Nvidia’s software ecosystem. Users can develop and test AI models on their desktop systems running Nvidia’s DGX OS, then scale these models to Nvidia DGX Cloud, on-premises data centers, or accelerated cloud instances. This integration is facilitated by Nvidia AI Enterprise, which provides a consistent environment from development through to production, ensuring that AI applications can be scaled efficiently without compromising on performance or compatibility.
The software suite available for Project DIGITS includes Nvidia’s comprehensive AI software stack – from SDKs and orchestration tools to frameworks like PyTorch and libraries like Nvidia NeMo and RAPIDS. Developers can leverage these tools for everything from model prototyping to deploying agentic AI applications using Nvidia Blueprints and NIM microservices.
For those looking to transition their AI development from research to real-world application, Nvidia AI Enterprise provides enterprise-grade security, support, and updates, ensuring that the journey from experimentation to production is smooth and secure.
Project DIGITS will become available in May 2025, starting at a price point of $3,000, making advanced AI computing accessible to a broader audience. This move by Nvidia not only underscores their commitment to advancing AI technology but also aims to foster a new wave of innovation by placing powerful AI tools in the hands of more creators, researchers, and students worldwide.
WallStreetPit does not provide investment advice. All rights reserved.
Leave a Reply