Dell DGX GB10 128G Ultra High Speed Response Memory Low Delay SSD Scientific Research AI Development Workstation

Product Description

Product Presentation Hardware View
Technical Specifications
System Architecture Chipset: GB10 Grace Blackwell Superchip. Combines a 20-core Arm CPU with a Blackwell GPU via NVLink-C2C interconnect. Coherent unified memory architecture.
CPU 20-core Arm v9 processor (10× Cortex-X925 high-performance cores, 10× Cortex-A725 energy-efficient cores)
GPU Blackwell architecture with 6,144 CUDA cores, 5th-generation Tensor Cores (1,000 AI TOPS FP4), and 4th-generation RT Cores.
Video Engines 1× NVENC (9th gen), 1× NVDEC (5th gen)
System Memory 28GB LPDDR5X unified memory (256-bit interface, 273GB/s bandwidth). Supports models up to 200B parameters locally.
Storage 4TB NVMe M.2 SSD (self-encrypting)
Networking ConnectX-7 Smart NIC (200Gbps RDMA for clustering), 10GbE RJ45 port, Wi-Fi 7 (802.11be) + Bluetooth 5.3
Interfaces 4× USB4 Type-C (40Gbps), 1× HDMI 2.1a (8K video output), Multichannel Audio
Power Consumption 170W typical operation; up to 224W peak under heavy load
Cooling Passive airflow design optimized for silent operation
Physical Specs Dimensions: 150mm(L) x 150mm(W) x 50.5mm(H); Weight: 1.2kg; Form Factor: FHFL dual-slot
Operating System DGX OS (Ubuntu 22.04-based) with pre-installed AI Enterprise software stack (PyTorch, TensorFlow, RAPIDS)
Model Capabilities Inference: Up to 200B parameters (LLaMA 2, DeepSeek-R1); Fine-tuning: Up to 70B parameters
Security

Cyber Resilient Architecture

Zero Trust IT environment security integrated into every phase. Silicon-based root of trust anchors end-to-end boot resilience.

Operations

Autonomous Collaboration

Increase efficiency with automated management. Simplify and centralize one-to-many management with enterprise consoles.

Sustainability

Sustainability

Designed with recycled materials and innovative energy efficiency options to reduce carbon footprint and operation costs.

DGX Spark Professional Deployment

Designed for 24x7 enterprise data center operations, DGX Spark features enterprise-grade components and energy-efficient hardware optimized for deployment at scale.

It supports Secure Boot with Hardware Root of Trust technology, providing an extra layer of security, and meets NEBS Level 3 standards for modern data center requirements. The passively cooled, energy-efficient design fits a wide range of Tesla-Certified systems and is easy to integrate into existing infrastructure.

Data Center Integration
Detail 1 Detail 2 Detail 3
Detail 4 Detail 5 Detail 6
Final Overview
Frequently Asked Questions
What core architecture does the DGX Spark utilize?
It utilizes the GB10 Grace Blackwell Superchip, which integrates a 20-core Arm CPU with a Blackwell GPU using NVLink-C2C high-speed interconnects.
What are the local AI model inference capabilities?
The system supports local inference for models up to 200 billion parameters, such as LLaMA 2 and DeepSeek-R1, and can scale further through clustering.
What networking options are available for high-performance clustering?
The system includes a ConnectX-7 Smart NIC supporting 200Gbps RDMA, alongside a 10GbE RJ45 port and Wi-Fi 7/Bluetooth 5.3 connectivity.
Is the hardware designed for secure enterprise environments?
Yes, it features a Cyber Resilient Architecture with Hardware Root of Trust, Secure Boot, ECC memory protection, and ARM TrustZone for confidential computing.
What software stack comes pre-installed with the system?
It comes with DGX OS (Ubuntu-based) and the AI Enterprise software suite, including cuDNN, TensorRT, and support for PyTorch and TensorFlow frameworks.
What is the typical power consumption and cooling requirement?
Typical operation consumes 170W, with a peak of 224W. The system is designed with passive airflow for silent, energy-efficient operation in diverse environments.

Related Products