Agentic AI
Build, run and manage always-on AI agents privately from your desk, from single-agent prototypes to multi-agent workflows.
NVIDIA GB10 Grace Blackwell Superchip
DGX Spark 128GB puts a petaFLOP of AI performance and 128 GB of unified memory on your desk, enough to run models up to 200 billion parameters locally and privately.
Overview
Powered by the NVIDIA GB10 Grace Blackwell Superchip, DGX Spark combines a 20-core Arm CPU and a Blackwell GPU that share one coherent pool of memory. It ships with the NVIDIA AI software stack, so agents and large models run on your own hardware instead of rented cloud tokens.

Workloads
Built for AI developers, researchers and data scientists who need data-center architecture at the desk.
Build, run and manage always-on AI agents privately from your desk, from single-agent prototypes to multi-agent workflows.
Develop, test and validate models and AI-augmented applications before migrating to DGX Cloud or the data center.
Customize pre-trained models of up to 70 billion parameters using the full 128 GB of unified memory.
Run and validate models up to 200 billion parameters with FP4 support on fifth-generation Tensor Cores.
Accelerate large analytics and machine-learning pipelines with 1 PFLOP of parallel throughput.
Develop robotics, smart-city and vision solutions with NVIDIA Isaac, Metropolis and Holoscan.
Specifications
Scale
Connect up to four DGX Spark units with ConnectX-7 to pool memory for larger models and faster inference.
Get started
Explore step-by-step playbooks and find where to buy on NVIDIA's official pages.