NVIDIA DGX Spark: World's Smallest AI Supercomputer Ships

NVIDIA Revolutionizes AI Development with Desktop Supercomputer

NVIDIA has officially begun shipping the DGX Spark, heralded as the world's smallest AI supercomputer, marking a significant milestone in democratizing artificial intelligence development. The compact desktop system, which started shipping this week, delivers unprecedented computing power directly to developers' desks, potentially transforming how AI research and development is conducted globally.

Powerful Performance in Compact Form

The DGX Spark represents a new class of computer specifically designed to address the growing computational demands of modern AI workloads. 'AI workloads are quickly outgrowing the memory and software capabilities of the PCs, workstations and laptops millions of developers rely on today,' NVIDIA noted in their announcement. The system delivers a remarkable one petaflop of AI performance and 128GB of unified memory in a form factor small enough for any office or lab environment.

Built on NVIDIA's cutting-edge Grace Blackwell architecture, the DGX Spark integrates the company's complete AI platform - including GPUs, CPUs, networking, CUDA libraries and the full NVIDIA AI software stack. The system is powered by the NVIDIA GB10 Grace Blackwell Superchip, which features 20 Arm CPU cores and 6,144 CUDA cores, all manufactured on TSMC's advanced 3nm process according to TechPowerUp.

Historical Connection to AI Revolution

In a symbolic gesture connecting past and present AI breakthroughs, NVIDIA founder and CEO Jensen Huang personally delivered one of the first DGX Spark units to Elon Musk at SpaceX's Starbase facility in Texas. 'In 2016, we built DGX-1 to give AI researchers their own supercomputer. I hand-delivered the first system to Elon at a small startup called OpenAI - and from it came ChatGPT, kickstarting the AI revolution,' Huang recalled during the delivery.

This delivery echoes the historic 2016 moment when Huang delivered the first DGX-1 supercomputer to OpenAI, which ultimately contributed to the development of ChatGPT and the subsequent AI boom. The DGX-1, as detailed on Wikipedia, was NVIDIA's first dedicated AI supercomputer system that helped establish the company's leadership in AI computing.

Technical Specifications and Capabilities

The DGX Spark's technical capabilities are impressive for its compact size. According to StorageReview, the system can run inference on AI models with up to 200 billion parameters and fine-tune models of up to 70 billion parameters locally. This eliminates the need for developers to shift work to cloud or data center environments for many advanced AI tasks.

The system features NVIDIA ConnectX-7 200 Gb/s networking and NVLink-C2C technology, providing five times the bandwidth of fifth-generation PCIe. The unified 128GB memory architecture enables seamless CPU-GPU coherency, a critical feature for handling large AI models efficiently.

Ecosystem and Partner Support

NVIDIA has secured broad ecosystem support for the DGX Spark, with major technology companies including Google, Meta, Microsoft, Hugging Face, and others already testing and optimizing their tools and models for the platform. 'DGX Spark allows us to access peta-scale computing on our desktop,' said Kyunghyun Cho, professor of computer and data science at the NYU Global Frontier Lab. 'This new way to conduct AI research and development enables us to rapidly prototype and experiment with advanced AI algorithms and models - even for privacy- and security-sensitive applications, such as healthcare.'

The system comes pre-installed with NVIDIA's complete AI software stack, enabling developers to start working immediately on AI projects. This includes access to NVIDIA AI ecosystem tools, models, libraries, and NVIDIA NIM microservices.

Availability and Pricing

Starting October 15, 2025, the DGX Spark is available for order directly from NVIDIA.com with a price point of $3,999. Partner systems will also be available from Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo, and MSI, as well as through Micro Center stores in the United States and NVIDIA channel partners worldwide.

The compact device measures just 150x150x50.5mm and weighs approximately 1.2kg, making it one of the most portable high-performance computing solutions available. Despite its small size, it consumes only 240W of power, making it suitable for office environments without specialized power requirements.

As AI continues to evolve at a rapid pace, the DGX Spark represents NVIDIA's commitment to putting powerful AI computing capabilities directly into the hands of developers worldwide, potentially accelerating the next wave of AI innovation across industries from healthcare to autonomous systems.