Nvidia’s early AI strategy has tended to be hyper-focused on hyperscalers with the intention of helping these global internet and cloud giants build out massive AI data centers and other infrastructure around the world. But Nvidia CEO Jensen Huang’s latest Computex keynote this week in Taiwan suggested the company is making more of an effort to extend its AI appeal to the enterprise market.
That was evident in Huang’s public unveiling of Nvidia’s new DGX Spark and DGX Station devices, which he described as “personal AI supercomputers” that are “direct descendants” of Nvidia’s original DGX-1 platform, while still relying on their own built-from-the ground-up capabilities that reflect current AI needs
The DGX Spark is a 128 GB laptop capable of 1 petaflop AI performance, while the DGX Station is a developer desktop that aims for 20 petaflops AI performance. DGX Spark is built on the GB10 Grace Blackwell Superchip and DGX Station leverages Nvidia’s GB300 Grace Blackwell Ultra Desktop Superchip, according to data provided by Dion Harris, senior director of HPC and AI factory at Nvidia, during a press briefing prior to Huang’s keynote.
DGX Station also supports multi-Instance GPU technology to partition into as many as seven instances — each with its own high-bandwidth memory, cache and compute cores — serving as a personal cloud for data science and AI development teams. It also uses AI software similar to that used in larger-scale AI factories.
In addition to these machines, Nvidia also announced with its RTX Pro server for the enterprise, built with RTX PRO 6000 Blackwell Server Edition GPUs,
“Blackwell meets the needs of the hyperscaler, but it's also focused on the enterprise,” Harris said. “We’re in the midst of a trillion-dollar IT infrastructure transition to enterprise AI factories.”
But many enterprise IT teams are going to need a pretty detailed roadmap to help them make this transition. In addition to the new DGX and RTX devices Nvidia announced, the company also unveiled Enterprise AI Factory and AI Data Platform reference designs aimed at giving enterprise a starting point and some guideposts for the AI strategies. These designs encompass deployment of RTX PRO Blackwell servers, Spectrum-X Ethernet networking, BlueField DPUs, Nvidia-certified storage systems and Nvidia AI Enterprise software, which can be put into action to accelerate product design and engineering simulation applications, as well as a quickly growing catalog of AI-enabled business systems and teams of digital AI agents.
DGX Spark will be available from Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo and MSI, as well as global channel partners, starting in July. , while DGX Station is expected to be available from ASUS, Dell Technologies, GIGABYTE, HP and MSI later this year. The company has not yet announced an availability date for the RTX Pro.