At the Consumer Electronics Show (CES) in Las Vegas, NVIDIA has taken a bold step toward democratizing artificial intelligence (AI) with the launch of Project DIGITS, a personal AI supercomputer designed to bring cutting-edge AI capabilities to researchers, data scientists, and students worldwide. Powered by the groundbreaking NVIDIA GB10 Grace Blackwell Superchip, Project DIGITS delivers a staggering petaflop of AI computing performance, enabling users to prototype, fine-tune, and run large AI models directly from their desktops. This innovation marks a pivotal moment in AI development, as it places the power of a supercomputer within reach of millions of developers, potentially accelerating the pace of AI innovation across industries.

A petaflop on your desk

The centerpiece of Project DIGITS is the NVIDIA GB10 Superchip, a system-on-a-chip (SoC) built on the Grace Blackwell architecture. The GB10 combines an NVIDIA Blackwell GPU with fifth-generation Tensor Cores and a high-performance NVIDIA Grace CPU, connected via NVLink-C2C chip-to-chip interconnect. This integration delivers up to 1 petaflop of AI performance at FP4 precision, all while operating on standard electrical power.

“AI will be mainstream in every application for every industry. With Project DIGITS, the Grace Blackwell Superchip comes to millions of developers,” said Jensen Huang, founder and CEO of NVIDIA. “Placing an AI supercomputer on the desks of every data scientist, AI researcher, and student empowers them to engage and shape the age of AI.”

The GB10 Superchip’s power efficiency and performance are further enhanced by its collaboration with MediaTek, a leader in Arm-based SoC designs. Each Project DIGITS unit features 128GB of unified memory and up to 4TB of NVMe storage, enabling developers to run AI models with up to 200 billion parameters. For even larger models, two Project DIGITS systems can be linked via NVIDIA ConnectX networking to handle models with up to 405 billion parameters.

Seamless scaling from desktop to cloud

One of the most compelling features of Project DIGITS is its ability to bridge the gap between local development and large-scale deployment. Developers can prototype and fine-tune models on their desktop systems using the Linux-based NVIDIA DGX OS, then seamlessly deploy them on NVIDIA DGX Cloud, accelerated cloud instances, or data center infrastructure. This flexibility ensures that AI projects can scale efficiently without requiring significant changes to the underlying architecture.

The system also provides access to NVIDIA’s extensive AI software ecosystem, including the NVIDIA NeMo framework for fine-tuning models, RAPIDS libraries for data science acceleration, and popular frameworks like PyTorch and Jupyter notebooks. Additionally, developers can leverage NVIDIA Blueprints and NIM microservices to build agentic AI applications, with enterprise-grade support available through the NVIDIA AI Enterprise license.

AI innovation for the masses

The launch of Project DIGITS is part of NVIDIA’s broader strategy to address the growing demands of generative AI and large language models (LLMs). The Grace Blackwell architecture, named in honor of mathematician David H. Blackwell, represents a significant leap forward in AI computing. With 208 billion transistors—more than 2.5 times the number in NVIDIA’s previous Hopper GPUs—the Blackwell architecture is the largest GPU ever built, capable of delivering 20 petaFLOPS of compute power.

The architecture’s second-generation Transformer Engine and advanced NVLink interconnects enable unprecedented performance for training and inference, supporting models with trillions of parameters. These innovations are critical as industries ranging from healthcare to automotive increasingly rely on generative AI for real-time applications.

Project DIGITS will be available starting in May 2024, with prices beginning at $3,000. This affordability, combined with its groundbreaking performance, positions Project DIGITS as a game-changer for AI development. By bringing super computing power to individual developers, NVIDIA is not only democratizing access to AI but also fostering a new wave of innovation that could reshape industries and redefine what’s possible in the age of AI.

Sources include:

NVIDIANews.nvidia.com

Resources.nvidia.com

NVIDIA.com