Nvidia has started shipping the DGX Spark, a desktop AI supercomputer that integrates the company’s Grace Blackwell architecture. The system combines GPUs, CPUs, networking, and AI software to support local AI workloads. Hardware partners including Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo, and MSI are also releasing their own systems based on the DGX Spark.
According to Nvidia, the DGX Spark can deliver up to 1 petaflop of AI performance and includes 128 GB of unified memory. The system comes with preinstalled software for AI model training and inference. Orders for the DGX Spark are scheduled to open on October 15, 2025, through Nvidia’s website, with partner systems becoming available globally.
The DGX Spark is priced at $4,000, but its 273 GB/s memory bandwidth limits its throughput for production inference workloads. This positions the system more for prototyping and experiments than for full-scale deployment. Benchmarks indicate the DGX Spark is approximately four times slower than the RTX Pro 6000 Blackwell workstation GPU and also performs behind the RTX 5090 on large models due to its bandwidth constraints.
The unit features a compact chassis that maintains stable thermals under load. It draws around 170 W of power from an external USB-C source, a configuration that could complicate office rollouts. To connect two DGX Spark units for processing 405-billion-parameter models, additional ConnectX-7 200 GbE hardware is required. This adds to the $3,999 base price, making the total cost of ownership less straightforward than public-cloud GPU options.
The NYU Global Frontier Lab noted the DGX Spark’s suitability for privacy-sensitive work in healthcare, which creates opportunities for managed services covering procurement, HIPAA-compliant implementation, and ongoing security. The system supports fine-tuning for models up to 70 billion parameters, which is suitable for educational institutions and smaller biotech firms seeking local model customization without cloud data exposure. This has opened a market for turnkey AI lab setups.
Nvidia’s partner ecosystem, which includes Dell, HP, Lenovo, and ASUS, provides extensive channel reach. This allows integrators to bundle services such as installation, training, and support for organizations that do not have in-house AI expertise. In other recent developments, Nvidia’s CEO highlighted the company’s first direct partnership with OpenAI.
