NVIDIA to release two AI personal computers The company said at a GPU technology conference on Tuesday that it was in a new brand called DGX (formerly Project Digits) to help researchers, scientists and developers.
The company said the smaller DGX Spark and desktop-sized DGX stations will feature NVIDIA’s Blackwell Ultra platform. These computers are targeted at developers, researchers, robotics developers, data scientists and students to adjust AI models locally. If the power supply on the machine is not enough, power acceleration development can also be used with the NVIDIA DGX cloud. The price of DGX Spark is $4,000, and Booking is now open. The price and availability of the DGX station will arrive later this year, and the device will be manufactured by NVIDIA’s hardware partners, including Asus, Boxx, Dell, HP, HP, Lambda and Supermicro.
According to the company’s press release, DGX Spark will use NVIDIA GB10 Grace Blackwell SuperChip and fifth-generation tensor cores, “AI Compute that can be used for fine-tuning and reasoning up to 1,000 trillion operations per second.” NVIDIA says the machine can handle the latest inference models, such as DeepSeek R1.
The DGX station will use the more powerful GB300 Grace Blackwell Ultra desktop super chip with 784GB of coherent memory space for large-scale training and inference.
Watch the following: Watch NVIDIA’s GTC 2025 Keynote: All Highlights in 16 Minutes
As AI has become an integral part of product development, more and more companies and individuals are looking for on-site solutions. When using the service from OpenAI, Google, Anthropic, and others, the costs associated with using AI on their servers. After a lot of use, these costs quickly add up. For some, like researchers and creators who are constantly iterating on projects, having equipment that can do all the leg work on the spot can save a lot of money. In addition, there is less concern for slowing down servers due to high loads. Data stored on local machines is also more secure. For financial institutions or hospitals that prefer sensitive data on site, local AI computers are preferred.
