Genuine NVIDIA through authorized distribution — NDAA §889 screened before checkout.
SKU
Category
Desktop Computers
Overview
Desktop Agent Computer
Powered by the NVIDIA GB10 Grace Blackwell Superchip, NVIDIA DGX Spark™ is a complete platform for local autonomous agents. Large local memory, powerful AI compute and the NVIDIA AI software stack enable agents and large models to run locally, reducing the need for cloud-based token generation resources. With a compact, power efficient design, DGX Spark is built to run always-on agent workloads - right from the desktop.
NVIDIA DGX Spark™ is part of a new class of supercomputers designed from the ground up to build and run AI. Powered by the NVIDIA GB10 Grace Blackwell Superchip and based on the NVIDIA Grace Blackwell architecture, NVIDIA DGX Spark delivers up to 1 petaFLOP¹ of AI performance to power large AI workloads. The GB10 Superchip uses NVIDIA NVLink™-C2C technology to deliver a CPU+GPU coherent memory model with 5x the bandwidth of PCIe Gen 5 and combined with 128 GB of unified system memory, developers can experiment, fine-tune, or build agents using the latest generation of reasoning and MoE AI models of up to 200B parameters.² Plus, NVIDIA ConnectX networking can connect up to four DGX Spark systems to work with models up to 700B parameters.²,³
NVIDIA DGX Spark is powered by NVIDIA DGX™ OS with Ubuntu Linux and comes preconfigured with the latest NVIDIA AI software stack. With NVIDIA Developer Program access to NVIDIA NIM™ and NVIDIA Blueprints, developers can hit the ground running using common tools such as PyTorch, Jupyter, and Ollama to prototype, fine-tune, and run inference locally. Work can then be easily moved to DGX Cloud or any accelerated data center or cloud infrastructure. Large local memory, strong compute performance, and power-efficient design make DGX Spark ideal for desktop autonomous agent development and deployment. NVIDIA OpenShell™ provides an open source runtime to build always-on autonomous, self-evolving agents more safely on DGX Spark.
1. Theoretical FP4 TOPS using the sparsity feature. | 2. Using FP4 precision models. | 3. Connecting four DGX Spark systems requires external switch
Highlights
Built on NVIDIA GB10 Grace Blackwell Superchip
NVIDIA Blackwell GPU with fifth-generation Tensor Core technology
NVIDIA Grace CPU with 20-core high-performance Arm architecture
Up to 1 petaFLOP¹ of AI performance using FP4
128 GB of coherent, unified system memory
Support for up to 200 billion parameter² models
NVIDIA ConnectX™ networking to link up to four systems to work with models up to 700 billion parameters³
4 TB of NVMe storage
Compact desktop form factor
Specifications
Architecture
NVIDIA Grace Blackwell
GPU
Blackwell Architecture
CPU
20-core Arm, 10 Cortex-X925 + 10 Cortex-A725 Arm
CUDA Cores
6,144
Tensor Cores
5th Generation
RT Cores
4th Generation
Tensor Performance
Up to 1 PFLOP FP4
System Memory
128 GB LPDDR5x, coherent unified system memory
Memory Interface
256-bit
Memory Bandwidth
273 GB/s
Storage
4 TB NVME.M2 with self-encryption
USB
4x USB TypeC
Ethernet
1x RJ-45 connector 10 GbE
NIC
ConnectX-7 NIC @ 200 Gbps
Wi-Fi
WiFi 7
Bluetooth
BT 5.4
Audio-output
HDMI multichannel audio output
Power Supply
240 Watts
GB10 TDP
140 W
Display Connectors
1x HDMI 2.1a, Up to 3x DisplayPort over USB-C (DP Alt Mode)
