At the GTC 2025 event, Nvidia unveiled two groundbreaking “personal AI supercomputers” that promise to bring cutting-edge AI capabilities directly to desktops. These new systems, the DGX Spark and DGX Station, are powered by Nvidia’s innovative Grace Blackwell chip platform and are designed to cater to a wide range of users, from developers and researchers to students.
Key Features of the DGX Spark and DGX Station
The following are the key features of DGX Spark and DGX Station.
DGX Spark
This system is touted as the world’s smallest AI supercomputer, delivering an impressive 1,000 trillion operations per second, thanks to the powerful GB10 Grace Blackwell Superchip. Its compact design makes it an ideal choice for individuals and teams looking to prototype and fine-tune AI models locally. It is particularly well-suited for researchers, developers, and students who need a portable yet powerful solution for their AI needs.
DGX Station
A more powerful solution, the DGX Station is equipped with the GB300 Grace Blackwell Ultra Desktop Superchip and boasts a massive 784GB of memory. This configuration enables large-scale AI training and inferencing, making it a powerful tool for professional data scientists and AI practitioners. It also includes the ConnectX-8 SuperNIC, which supports networking speeds of up to 800Gb/s, allowing seamless data transfer and optimal performance.
Both systems are designed to offer flexibility, enabling users to transition effortlessly between local AI workflows and cloud-based platforms. This versatility ensures that both the DGX Spark and DGX Station are invaluable for developers and data scientists who require both local and cloud-based solutions for their AI models.
Availability and Pricing
- DGX Spark: Available for reservation immediately, this system is expected to have a base configuration starting at around $3,000. However, official pricing details for both systems are yet to be confirmed.
- DGX Station: Expected to launch later this year, the DGX Station will be available through Nvidia’s manufacturing partners, including Asus, Boxx, Dell, HP, and Lenovo.
Nvidia’s Vision for the Future
During the keynote, Nvidia CEO Jensen Huang emphasized the transformative potential of these new systems, stating, “AI has transformed every layer of the computing stack. It stands to reason a new class of computers would emerge—designed for AI-native developers and to run AI-native applications.”
Huang also described the DGX Spark and DGX Station as “the computer of the age of AI,” adding that this new breed of AI supercomputers would revolutionize how AI models are created and deployed. With the introduction of these machines, Nvidia aims to shape the future of computing, enabling AI agents to become ubiquitous across industries.
