NVIDIA

NVIDIA DGX Spark Grace Blackwell AI Desktop Supercomputer (2 Pack)

A two-node bundle of NVIDIA's Grace Blackwell desktop supercomputer, pairing dual GB10 chips and 256GB of combined unified memory for local AI work

$9,499.99*
In Stock on Amazon.com
View on Amazon

*Price sourced from Amazon.com. Last updated:Jul 24, 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 2 Pack takes the company's personal AI desktop supercomputer and delivers it in pairs, bundled with connecting cables. At the heart of each unit sits the GB10 Grace Blackwell chip, the architecture NVIDIA positions for local AI development. The pitch is straightforward: supercomputer-class performance in a compact, energy-efficient enclosure that lives on a desk rather than in a server room. Each node is rated at up to 1 petaFLOP of AI performance and carries 128GB of unified memory, which brings the two-pack to a combined 256GB. This bundle is aimed at developers, researchers, and teams who want to fine-tune, run inference on, and analyze large models locally without leaning entirely on cloud resources.

Practical ownership centers on the NVIDIA AI software stack, which comes integrated so work built locally can be deployed to laptop, desktop, cloud, or data center targets. NVIDIA cites headroom for models up to 200 billion parameters at FP4 precision on a single unit's 128GB of unified memory, framing the hardware as a space to prototype, test, and iterate securely. The two-unit configuration suits buyers who expect to distribute work across nodes or who simply want a second machine on hand. Beyond the chip, memory, and performance figures, the listing stays light on specifics like storage capacity, connectivity, and physical dimensions, so anyone planning a build should confirm those details directly before committing.

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 (per unit) of unified memory. Empower rapid testing, validation, and iteration—driving innovation in a secure, high-performance setting.

Specifications

Brand
NVIDIA
Product
DGX Spark 2 Pack with Cable Bundle
Type
Personal AI desktop supercomputer
Chip
GB10 Grace Blackwell
AI Performance
Up to 1 petaFLOP per unit
Unified Memory
128GB per unit (256GB combined)
Max Model Size
Up to 200 billion parameters at FP4
Units Included
2 (plus cable bundle)
Software
Full NVIDIA AI software stack
Form Factor
Compact desktop

Frequently Asked Questions

Two DGX Spark units, each built around the GB10 Grace Blackwell chip, plus a cable bundle for connecting them, according to the product listing.
Each unit includes 128GB of unified memory, so the two-pack provides a combined 256GB for running and fine-tuning large local models.
NVIDIA states you can experiment with large models up to 200 billion parameters at FP4 precision on one unit's 128GB of unified memory.
The Grace Blackwell architecture is rated at up to 1 petaFLOP of AI performance per unit for local fine-tuning, inference, and analytics.
Yes. It runs the full NVIDIA AI software stack, so you can develop locally and deploy the same work to laptop, desktop, cloud, or data center resources.
Developers, researchers, and teams who want compact, energy-efficient desktop hardware for fine-tuning, inference, and analytics without relying solely on the cloud.