NVIDIA

NVIDIA DGX Spark Personal AI Desktop Supercomputer

Up to 1 petaFLOP of Grace Blackwell AI performance and 128GB of unified memory in a desktop built to run models up to 200 billion parameters

$4,679.00*
In Stock on Amazon.com
View on Amazon

*Price sourced from Amazon.com. Last updated:Jul 22, 2026.Price and availability are subject to change.

Affiliate Disclosure: Studio Supplies may earn a commission from qualifying purchases made through links on this page, at no additional cost to you. This helps support our editorial team.

Notice a mistake? Let Us Know

Overview

The NVIDIA DGX Spark is a personal AI supercomputer aimed squarely at people who build and run models for a living. At its core sits the GB10 Grace Blackwell chip, and NVIDIA rates the system for up to 1 petaFLOP of AI performance. That figure, paired with 128GB of unified memory, is what allows the machine to take on large-model fine-tuning, inference, and analytics locally. The intended buyer is a developer, data scientist, or research team that wants enterprise-scale capability without routing every job through the cloud. It is less a general desktop and more a dedicated laboratory for AI experimentation.

Practical ownership centers on the software as much as the silicon. The DGX Spark arrives with the full NVIDIA AI software stack, so the workflow is meant to be seamless: prototype on the desk, then deploy the same work to a laptop, cloud instance, or data center. The compact, energy-efficient design is a deliberate choice, letting the unit sit in an office rather than a server room. NVIDIA frames it as an augment to existing resources — a way to learn, prototype, test, and iterate quickly in a secure setting. With unified memory large enough to hold models up to 200 billion parameters at FP4, the appeal is rapid, self-contained iteration for those whose bottleneck is access to compute.

Key Features

Supercomputer performance directly to your desk in a compact, energy-efficient design, enabling enterprise-scale AI and high-performance computing right where you need it.

The power of Grace Blackwell architecture, delivering up to 1 petaFLOP of AI performance for local model fine-tuning, inference, and analytics, accelerating your time-to-solution.

Designed from the ground up to build and run AI, delivering seamless integration of the full NVIDIA AI software stack —so you can develop locally and deploy anywhere.

NVIDIA DGX Spark gives you the freedom to experiment, prototype, and innovate faster by augmenting laptop, desktop, cloud, or data center resources. With more power to learn, prototype, test, and innovate, NVIDIA DGX Spark delivers exceptional ROI for increased productivity.

Use NVIDIA DGX Spark to unlock new ideas and experiment with large models (up to 200 billion parameters at FP4) directly on your desktop with 128GB of unified memory. Empower rapid testing, validation, and iteration—driving innovation in a secure, high-performance setting.

Specifications

Brand
NVIDIA
Model
DGX Spark
Chip
GB10 Grace Blackwell
Architecture
Grace Blackwell
AI Performance
Up to 1 petaFLOP
Unified Memory
128GB
Max Model Size
Up to 200 billion parameters (FP4)
Software
Full NVIDIA AI software stack
Primary Use
Local fine-tuning, inference, analytics
Form Factor
Compact desktop

Frequently Asked Questions

NVIDIA states it can work with large models up to 200 billion parameters at FP4 precision, enabled by its 128GB of unified memory for local experimentation.
It uses the GB10 Grace Blackwell chip built on NVIDIA's Grace Blackwell architecture, rated for up to 1 petaFLOP of AI performance.
It targets developers, researchers, and enterprise teams who fine-tune, run inference, and analyze data locally, rather than mainstream desktop or gaming users.
Yes. It ships with the full NVIDIA AI software stack, so the intended workflow is to develop locally and deploy the same work to a laptop, cloud, or data center.
NVIDIA's pitch is freedom to prototype and iterate in a secure, high-performance setting on your own desk, augmenting laptop, desktop, cloud, or data center resources.