PNY NVIDIA DGX Spark 4TB
PNY NVIDIA DGX Spark 4TB
If you are interested in purchasing an AI solution or accelerator, please email us at [email protected]. We will discuss a specific solution with you and prepare a quote.
PNY NVIDIA DGX Spark 4TB mini computer
The world's smallest personal AI supercomputer
PNY NVIDIA DGX Spark 4TB is a groundbreaking miniature personal AI supercomputer, bringing the power of an entire AI data centre into a 150×150×50.5mm body for the very first time. At its core beats the NVIDIA GB10 Grace Blackwell superchip—a combination of a 20-core ARM processor and a Blackwell architecture GPU with 5th generation Tensor Cores, which together achieve a performance of up to 1 petaflop (PFLOP). A key advantage is its 128GB of shared unified CPU and GPU memory, interconnected by NVLink-C2C technology with 5× higher throughput than PCIe 5th generation. This makes it possible to load and run models with up to 200 billion parameters, such as Llama 3.1 70B or Gemma 3 27B, directly into memory and locally, without depending on the cloud or an external server. The 4TB M.2 NVMe encrypted storage, a ConnectX-7 network card with 200Gb/s InfiniBand, Wi-Fi 7, Bluetooth 5.4, and four 40Gb/s USB-C 4.0 ports ensure connectivity for every professional scenario—including linking two DGX Spark units into a mini cluster to work with models of up to 405 billion parameters. The pre-installed NVIDIA DGX OS, complete with CUDA libraries, Docker, and the Container Toolkit, enables immediate development without lengthy configuration. This system is based on Ubuntu Linux with integrated support for Ubuntu Pro Client for extended ESM security updates.
LPDDR5x / 128GB
Operating memory
Blackwell GB10
Chip architecture
NVIDIA ConnectX
High-speed network card
NVIDIA AI Software Stack
Complete software stack
240W
Power consumptionKey features of the PNY NVIDIA DGX Spark 4TB mini computer
- NVIDIA GB10 Grace Blackwell superchip with a 20-core ARM processor and a Blackwell architecture GPU with 5th generation Tensor Cores
- Up to 1 PFLOP of AI inference performance for local processing of the most demanding models
- 128GB of shared unified CPU and GPU memory connected by NVLink-C2C technology with 5× higher throughput than PCIe 5th generation
- The computer can load and run models with up to 200 billion parameters directly in memory
- The NVIDIA ConnectX-7 network card with 200Gb/s InfiniBand lets you link two DGX Spark systems into a mini cluster to work with models up to 405 billion parameters
- Encrypted 4TB M.2 NVMe storage provides ample space for large datasets, models and development projects with maximum security for stored data
- Pre-installed NVIDIA DGX OS with CUDA libraries, Docker and the Container Toolkit
- The system is based on Ubuntu Linux with integrated support for Ubuntu Pro Client for extended ESM security updates
- Supports PyTorch, TensorFlow and NVIDIA NIM microservices
- Wi-Fi 7, Bluetooth 5.4, 10Gb Ethernet, 4× USB-C 4.0 (40Gb/s) and HDMI provide complete connectivity in a body measuring 150×150×50.5mm and weighing only 1.2kg
Data centre performance in a miniature body
At the heart of the PNY NVIDIA DGX Spark 4TB is the GB10 Grace Blackwell superchip—a unique combination of a 20-core ARM processor (10× Cortex-X925 + 10× Cortex-A725) and a Blackwell architecture GPU with 5th generation Tensor Cores and 4th generation RT Cores. Both parts of the superchip are connected by NVLink-C2C technology, which provides 5× higher throughput than PCIe 5th generation and allows the CPU and GPU to share a single 128GB pool of unified LPDDR5X memory. The result is enormous performance that consumes a maximum of 240W and fits on any desk. This shared memory, without the traditional separation of GPU VRAM and system RAM, is the key advantage of the DGX Spark over standard desktop setups—where a GeForce RTX 5090 graphics card offers 32GB of memory, the DGX Spark makes four times that available for AI models, without needing to quantise models or split them between multiple devices.
A personal supercomputer that fits on any desk
The PNY NVIDIA DGX Spark 4TB proves that you no longer need a server room or a rack full of hardware to run the largest AI models. With dimensions of 150×150×50.5mm and a weight of 1.2kg, the DGX Spark fits on any desk next to a laptop—while offering performance that, just a few years ago, would have required an entire server full of graphics cards. Four USB-C 4.0 ports with speeds up to 40Gb/s, Wi-Fi 7, 10Gb Ethernet and an HDMI output ensure you can connect all your accessories without needing an external hub or docking station. The DGX Spark is designed for developers, scientists, and AI professionals who want the full power of a local personal AI supercomputer without depending on the cloud—right where they work.
Acceleration of all AI tasks in a compact body
The PNY NVIDIA DGX Spark 4TB is the ideal choice for developers, researchers, and data scientists who need full-fledged AI performance without compromise. Whether you're working on large language model inference, specialising pre-trained networks, or developing AI agents, the DGX Spark handles the full spectrum of AI tasks locally, quickly, and without reliance on cloud infrastructure.
Limitless prototyping and development of AI applications
The complete NVIDIA AI software stack provides developers with a full-featured platform for creating AI models, AI agents, and AI-enhanced applications – all locally, without relying on the cloud. Once a solution is ready for deployment or final fine-tuning, DGX Spark enables direct and seamless migration of workloads to NVIDIA DGX Cloud or other NVIDIA-accelerated infrastructure.
Fine-tune AI models with up to 70 billion parameters
Use the 128GB of unified memory in the PNY NVIDIA DGX Spark to fine-tune pre-trained models with up to 70 billion parameters right at your workstation. Training on your own data lets you specialise AI models for specific needs, industry data, or particular use cases – without sending sensitive data to the cloud or paying for expensive GPU instances. The result is a model tailored precisely to your scenario, created locally, securely, and under your complete control.
PNY NVIDIA DGX Spark 4TB facilitates data science at the highest level
The combination of 128GB of unified memory and 1 PFLOP of parallel throughput performance makes the PNY NVIDIA DGX Spark the ideal workstation for demanding data analytics and machine learning projects. Extensive datasets, complex computational models, and demanding training and analysis processes that previously required a powerful cloud cluster or dedicated server now run directly on your desk. Fast, local, and without waiting for remote infrastructure.
Inference for models with up to 200 billion parameters
The 5th generation Tensor Cores with FP4 format support achieve up to 1 PFLOP of performance and, combined with 128GB of system memory, enable you to run inference on state-of-the-art AI models with up to 200 billion parameters directly on your desk. Test, validate, and deploy models like Llama, Gemma, or Qwen locally and in real time – without cloud latency, shared infrastructure, or the risk of sensitive data leaks.
Application development for robotics, industry, and smart cities
The PNY NVIDIA DGX Spark is an exceptional platform for developing robotic systems, smart city solutions, and computer vision applications. The pre-installed NVIDIA Isaac framework for robotics, Metropolis for video analytics, and Holoscan for real-time data processing enable developers to fully harness the power of DGX Spark and rapidly prototype and deploy applications – locally, without dependence on cloud infrastructure, and with full support from the NVIDIA ecosystem.
The power of a personal AI supercomputer
Up to 1 petaflop for inference
Up to 200 billion parameters
Local inference without cloud dependency
Connect two systems
Mini cluster for models with up to 405 billion parameters
Ready to work straight away
Pre-installed NVIDIA AI software stack
Memory without compromise
128GB of unified CPU and GPU memory
Your own data
Fine-tune models with up to 70 billion parameters
Compact dimensions
150×150×50.5mm, 1.2kg
Secure local work
Full data control without sending to the cloudSpecifications
Use
Processor
GPU series/model
Graphics card
Hard Drive
Equipment
Colour and design
PSU
Memory
Outputs
Operating system
Package contents
AI
Security and management
External structure
AI specifications
GPU configuration
Cooling system
PNY Technologies Europe, ZAC du Phare 9 rue Joseph Cugnot 33708 Mérignac cedex, France https://www.pny.com/en-eu/support-footer/technical-support
Specifications can be changed without notice. Images are for illustrative purposes only.
