Nvidia has introduced two groundbreaking personal AI supercomputers, marking a significant step forward in the future of artificial intelligence computing. At the GTC 2025 conference, CEO Jensen Huang announced the launch of the DGX Spark and DGX Station. Both machines are powered by Nvidia’s latest Grace Blackwell chip platform and designed to meet the growing demands of AI workloads.
These advanced systems are built to help developers, researchers, and enterprises prototype, train, and deploy AI models more efficiently. With powerful computing capabilities packed into a desktop form, they bring the performance of large data centers directly to personal workstations.
DGX Spark: Compact Yet Extremely Powerful
The DGX Spark is the first in this new line of personal AI supercomputers. Built with Nvidia’s GB10 Grace Blackwell Superchip, the DGX Spark delivers a staggering 1,000 trillion operations per second. This level of computing power enables users to run complex AI models and fine-tune algorithms with ease, even on a compact system.
Designed for flexibility, the DGX Spark makes it easier for AI professionals to develop models at the edge, without relying solely on cloud resources. Its size and power make it ideal for those needing high-performance computing within a limited space. Whether running experiments or refining AI models, the DGX Spark offers unmatched capability for individual developers and smaller teams.
According to Nvidia, the DGX Spark is available for purchase immediately, providing access to its impressive computing performance right away.
DGX Station: Enterprise-Grade AI Power for Workstations
Alongside the DGX Spark, Nvidia also revealed the DGX Station, aimed at enterprises and large-scale AI projects. The DGX Station is equipped with Nvidia’s GB300 Grace Blackwell Ultra Desktop Superchip and offers an incredible 784GB of memory. This combination gives organizations the power to handle resource-intensive AI model training and deployment directly from their workstations.
By offering this level of computing performance in a desktop setup, the DGX Station provides businesses with a viable alternative to cloud-based AI development. Companies can now manage sensitive data on-premise while benefiting from lightning-fast processing speeds. For enterprises where speed, performance, and data privacy are top priorities, the DGX Station is a game-changer.
Nvidia announced that the DGX Station will be available later this year through manufacturing partners such as Asus, Boxx, Dell, HP, and Lenovo.
Shaping the Future of AI Computing
During his keynote, Jensen Huang emphasized that the age of AI requires a completely new approach to computing. Traditional systems can no longer meet the demands of AI agents, which are becoming an integral part of everyday business and personal applications.
“AI agents will be everywhere,” Huang stated. “How they operate, how enterprises run them, and how we manage these systems will fundamentally change. These new AI supercomputers are designed for that future.”
With these releases, Nvidia is positioning itself at the forefront of AI hardware innovation. Both the DGX Spark and DGX Station aim to empower businesses, researchers, and developers by offering powerful computing without the need for extensive server infrastructure or cloud dependency.
Key Highlights of Nvidia’s New AI Supercomputers
- DGX Spark delivers 1,000 trillion operations per second, powered by the GB10 Grace Blackwell Superchip.
- DGX Station offers 784GB of memory, using the GB300 Grace Blackwell Ultra Desktop Superchip.
- Both machines support AI model training, fine-tuning, and deployment at the edge.
- DGX Spark is available now, while DGX Station will be released later this year.
- Manufacturing partners include Asus, Boxx, Dell, HP, and Lenovo.
As AI becomes increasingly embedded in business operations and daily life, these powerful personal supercomputers are set to redefine how models are developed and deployed. Nvidia’s move reflects the growing need for localized AI computing power and the shift toward edge AI solutions.
