Trusted Partners

Powered by the NVIDIA GB10 Grace Blackwell Superchip, the NVIDIA DGX Spark™ provides up to one petaFLOP of FP4 AI performance within a compact, energy-efficient design. With 128 GB of memory and a preinstalled NVIDIA AI software stack, it enables developers to prototype, optimize, and deploy advanced reasoning AI models directly from their desktop.
Powered by the NVIDIA GB10 Grace Blackwell Superchip, the NVIDIA DGX Spark™ provides up to one petaFLOP of FP4 AI performance within a compact, energy-efficient design. With 128 GB of memory and a preinstalled NVIDIA AI software stack, it enables developers to prototype, optimize, and deploy advanced reasoning AI models directly from their desktop.


NVIDIA NemoClaw is an open-source platform that integrates privacy and security features into OpenClaw. It allows developers to execute secure, always-on, self-evolving AI agents on NVIDIA RTX PCs, DGX Stations, and DGX Spark systems with a single command.

NVIDIA NemoClaw is an open-source platform that integrates privacy and security features into OpenClaw. It allows developers to execute secure, always-on, self-evolving AI agents on NVIDIA RTX PCs, DGX Stations, and DGX Spark systems with a single command.

Leverage the power of the NVIDIA Grace Blackwell architecture, which delivers up to 1 petaFLOP of AI processing performance at FP4 precision.
Leverage the power of the NVIDIA Grace Blackwell architecture, which delivers up to 1 petaFLOP of AI processing performance at FP4 precision.


Equipped with 128 GB of coherent unified system memory, this workstation allows you to execute AI testing and development tasks with models containing up to 200 billion parameters directly on your desktop.

Equipped with 128 GB of coherent unified system memory, this workstation allows you to execute AI testing and development tasks with models containing up to 200 billion parameters directly on your desktop.

Integrated NVIDIA ConnectX™ networking technology provides the necessary bandwidth to link two NVIDIA DGX Spark systems, supporting complex AI models with up to 405 billion parameters.
Integrated NVIDIA ConnectX™ networking technology provides the necessary bandwidth to link two NVIDIA DGX Spark systems, supporting complex AI models with up to 405 billion parameters.


Take advantage of a complete software ecosystem designed for generative AI, featuring a robust collection of NVIDIA frameworks, tools, libraries, and pre-trained models such as NVIDIA NIM.

Take advantage of a complete software ecosystem designed for generative AI, featuring a robust collection of NVIDIA frameworks, tools, libraries, and pre-trained models such as NVIDIA NIM.

Utilizing the comprehensive NVIDIA AI software stack, the NVIDIA DGX Spark provides a dedicated environment for developers to build, test, and validate AI models, agents, and complex applications. Users can seamlessly perform final tuning and evaluation locally before transitioning projects to the NVIDIA DGX Cloud or other accelerated infrastructure.
Utilizing the comprehensive NVIDIA AI software stack, the NVIDIA DGX Spark provides a dedicated environment for developers to build, test, and validate AI models, agents, and complex applications. Users can seamlessly perform final tuning and evaluation locally before transitioning projects to the NVIDIA DGX Cloud or other accelerated infrastructure.


Boost the capabilities of pre-trained models using the high-performance NVIDIA DGX Spark. With 128GB of unified system memory, this system allows for the fine-tuning of sophisticated models containing up to 70 billion parameters, tailored to meet specific project requirements.

Boost the capabilities of pre-trained models using the high-performance NVIDIA DGX Spark. With 128GB of unified system memory, this system allows for the fine-tuning of sophisticated models containing up to 70 billion parameters, tailored to meet specific project requirements.

The integration of fifth-generation Tensor Cores with FP4 support offers up to 1 petaFLOP of AI computing power. Combined with 128GB of system memory, NVIDIA DGX Spark accelerates the inference of cutting-edge AI models, allowing for rapid testing, validation, and deployment.
The integration of fifth-generation Tensor Cores with FP4 support offers up to 1 petaFLOP of AI computing power. Combined with 128GB of system memory, NVIDIA DGX Spark accelerates the inference of cutting-edge AI models, allowing for rapid testing, validation, and deployment.


Experience enterprise-grade data science performance directly at your workstation. The NVIDIA DGX Spark leverages 128GB of unified memory and 1 petaFLOP of parallel throughput to expedite computationally intensive machine learning tasks and complex data analytics.

Experience enterprise-grade data science performance directly at your workstation. The NVIDIA DGX Spark leverages 128GB of unified memory and 1 petaFLOP of parallel throughput to expedite computationally intensive machine learning tasks and complex data analytics.

NVIDIA DGX Spark serves as an ideal development hub for robotics, computer vision, and smart city solutions. By supporting essential NVIDIA frameworks such as Isaac, Metropolis, and Holoscan, it enables developers to quickly architect and iterate on advanced edge computing applications.
NVIDIA DGX Spark serves as an ideal development hub for robotics, computer vision, and smart city solutions. By supporting essential NVIDIA frameworks such as Isaac, Metropolis, and Holoscan, it enables developers to quickly architect and iterate on advanced edge computing applications.


DGX Spark integrates with the NVIDIA AI Enterprise software suite, a cloud-native platform designed to accelerate production-grade AI. This ecosystem offers optimized performance, robust security, and professional support to help teams transition seamlessly from prototype to production-scale agentic AI.

DGX Spark integrates with the NVIDIA AI Enterprise software suite, a cloud-native platform designed to accelerate production-grade AI. This ecosystem offers optimized performance, robust security, and professional support to help teams transition seamlessly from prototype to production-scale agentic AI.
Everything you need to know about this product
The DGX Spark is powered by the NVIDIA GB10 Grace Blackwell Superchip, delivering up to 1 petaFLOP of FP4 AI computing performance via fifth-generation Tensor Cores.
With 128 GB of coherent unified system memory, the DGX Spark can run AI inference and testing on models with up to 200 billion parameters directly from the desktop.
Yes — the 128 GB of unified memory supports fine-tuning of models containing up to 70 billion parameters, allowing developers to tailor pre-trained models to specific project requirements locally.
Yes — integrated NVIDIA ConnectX networking technology provides the bandwidth to link two DGX Spark systems together, extending support to AI models with up to 405 billion parameters.
The DGX Spark ships with the complete NVIDIA AI software stack, including NVIDIA frameworks, tools, libraries, and pre-trained models such as NVIDIA NIM, ready for generative AI development.
Yes — developers can prototype, tune, and validate models locally on the DGX Spark before seamlessly transitioning projects to NVIDIA DGX Cloud or other accelerated infrastructure.
Yes — the DGX Spark supports NVIDIA Isaac, Metropolis, and Holoscan frameworks, making it a capable development hub for robotics, computer vision, and smart city edge applications.
Yes — NVIDIA NemoClaw is an open-source platform that enables developers to run secure, always-on, self-evolving AI agents on DGX Spark systems with a single command.
Yes — DGX Spark integrates with the NVIDIA AI Enterprise software suite, a cloud-native platform offering optimized performance, robust security, and professional support for production-grade agentic AI.
The NVIDIA DGX Spark is a compact, energy-efficient desktop workstation, designed to deliver up to 1 petaFLOP of AI performance from a personal workstation footprint.
AI-Generated Content: These FAQs are generated by an automated AI system and may contain inaccuracies or omissions. Evetech makes no warranties, express or implied, regarding their completeness or reliability. For definitive information, please refer to the manufacturer's official specifications.





Please Note: Product images are for illustrative purposes only and may differ from the actual product.
Hand-pick every part with live stock, compatibility, and instant pricing.
Try the new builderHand-pick every part with live stock, compatibility, and instant pricing.
Try the new builderNVIDIA DGX Spark Mini PC





