Nvidia announces a version of DGX Spark with 64 GB of unified memory for $4,999, or $1,000 more than the 128 GB version at launch

Local AI development becomes more attainable with a new, lower-cost hardware option.

Nvidia is launching a new configuration of its DGX Spark personal AI supercomputer, the DGX Spark 64GB, available starting October 23rd for $4,999. This version offers 64GB of unified memory, a reduction from the previously launched 128GB model. The new SKU is designed to make local AI development more accessible and affordable, allowing developers to run models up to 100 billion parameters entirely on device. It comes pre-loaded with DGX OS and the NVIDIA AI software stack. The DGX Spark 64GB will be available through manufacturer partners including Acer, ASUS, Dell, Gigabyte, HP, and MSI. The system supports clustering two units together using the NVIDIA Sync Cluster Assistant to pool memory and increase performance, enabling support for up to 200 billion parameter models.

The introduction of the 64GB DGX Spark at a lower price point signifies Nvidia's strategy to broaden access to its AI development hardware beyond high-end users. By offering a more affordable entry point, Nvidia aims to capture a larger segment of developers and researchers who are increasingly looking to run AI models locally for privacy and cost-efficiency. This move acknowledges the trend of AI models shrinking and becoming more feasible for on-device deployment, directly catering to the evolving needs of the AI development community.

This tiered offering allows developers to scale their AI infrastructure incrementally, starting with the 64GB model and expanding to a clustered 128GB setup as their project requirements grow. The seamless clustering capability, facilitated by NVIDIA Sync Cluster Assistant, suggests a future where personal AI supercomputing can adapt fluidly to increasing computational demands. This approach not only reduces upfront investment but also provides a clear upgrade path, potentially accelerating adoption of local AI development platforms across various industries.