The NVIDIA DGX Spark is a compact, powerful AI supercomputer designed for researchers, data scientists, robotics developers, and students. It was previously known as Project Digits and has been rebranded as part of NVIDIA's latest offerings in AI computing. Here are some key features and details about the DGX Spark:
Key Features of DGX Spark
- Architecture: Powered by the NVIDIA GB10 Grace Blackwell Superchip, which includes a Blackwell GPU with 5th-generation Tensor Cores[2][5].
- CPU: Equipped with 20 Arm CPU cores, specifically 10 Cortex-925 and 10 Cortex-725 cores[4].
- Memory: Offers up to 128GB of LPDDR5X unified memory, with a memory bandwidth of 273 GB/s[2][4].
- Storage: Configurable with up to 4TB of NVMe SSD storage[2][5].
- Connectivity: Includes Wi-Fi 7, Bluetooth 5.3, and four USB Type-C ports with speeds up to 40Gb/s[4].
- Power Consumption: Consumes up to 170W of power[4].
- Operating System: Runs on NVIDIA's custom DGX OS, which is based on Linux[3][4].
- Performance: Delivers up to 1 petaflop (1,000 TOPS) of AI performance at FP4 precision[2][5].
Availability and Pricing
- Pricing: Starts at $3,000[2][5].
- Availability: Reservations are open, with shipping expected to begin in summer 2025. It will also be available from partners like ASUS, Dell, HP, and Lenovo[2][5].
Purpose and Target Audience
The DGX Spark is designed to democratize access to advanced AI computing, allowing users to prototype, fine-tune, and deploy large AI models locally or in the cloud. It is particularly suited for AI developers, researchers, and students who need powerful AI capabilities in a compact form factor[1][5].
Citations:
[1] https://nvidianews.nvidia.com/news/nvidia-announces-dgx-spark-and-dgx-station-personal-ai-computers
[2] https://beebom.com/nvidia-project-digits-rebranded-to-dgx-spark-dgx-station-announced/
[3] https://www.pcmag.com/news/what-is-nvidias-dgx-station-a-new-specialized-desktop-line-for-ai-work
[4] https://www.reddit.com/r/LocalLLaMA/comments/1jedlum/dgx_sparks_nvidia_digits/
[5] https://www.theverge.com/news/631957/nvidia-dgx-spark-station-grace-blackwell-ai-supercomputers-gtc
[6] https://marketplace.nvidia.com/en-gb/developer/dgx-spark/
[7] https://www.youtube.com/watch?v=krBh0Von-2A
[8] https://www.theregister.com/2025/03/18/gtc_frame_nvidias_budget_blackwell/
[9] https://techcrunch.com/2025/03/18/nvidia-announces-two-personal-ai-supercomputers/
