AI Workstation // NVIDIA
NVIDIA DGX Spark AI Supercomputer – 128GB Unified Memory, 4TB NVMe SSD
Price inclusive of all taxes
Technical specification
NVIDIA DGX Spark – Personal AI Supercomputer
Bring advanced AI development to your desktop with the NVIDIA DGX Spark, a compact AI supercomputer powered by the NVIDIA GB10 Grace Blackwell Superchip. Designed for developers, researchers, and AI professionals, DGX Spark provides a powerful platform for running large AI models locally, experimenting with generative AI, and developing intelligent applications without relying entirely on cloud infrastructure.
With 128GB of coherent unified memory, up to 1 PFLOP of theoretical FP4 AI performance, and 4TB of NVMe storage, DGX Spark combines substantial memory capacity and AI computing power in an exceptionally compact form factor.
Key Features
- NVIDIA GB10 Grace Blackwell Superchip: Combines CPU and GPU computing in an integrated AI-focused platform.
- 128GB Unified Memory: High-capacity shared memory for supported AI models, development workloads, and data processing.
- Up to 1 PFLOP FP4 AI Performance: Designed to accelerate supported AI inference and development workloads.
- 4TB NVMe Storage: High-capacity internal storage with self-encryption for models, datasets, and development projects.
- 20-Core Grace CPU: Features 10 Arm Cortex-X925 cores and 10 Arm Cortex-A725 cores.
- Blackwell Tensor Cores: Fifth-generation Tensor Cores and fourth-generation RT Cores.
- High-Speed Networking: ConnectX-7 networking with up to 200Gbps connectivity, plus 10GbE Ethernet.
- Compact Design: A small desktop system measuring approximately 150 × 150 × 50.5 mm.
- NVIDIA DGX OS: A purpose-built software environment for AI development and deployment.
Specifications
| Product | NVIDIA DGX Spark |
| AI Superchip | NVIDIA GB10 Grace Blackwell |
| Architecture | NVIDIA Grace Blackwell |
| CPU | 20-core Arm CPU: 10 Cortex-X925 + 10 Cortex-A725 |
| Unified Memory | 128GB LPDDR5x |
| Memory Interface | 256-bit |
| Memory Bandwidth | 273GB/s |
| AI Performance | Up to 1 PFLOP FP4, theoretical |
| Tensor Cores | Fifth Generation |
| RT Cores | Fourth Generation |
| Storage | 4TB NVMe M.2 SSD with self-encryption |
| Operating System | NVIDIA DGX OS |
| Ethernet | 10GbE RJ-45 |
| Networking | ConnectX-7 Smart NIC with two QSFP network ports |
| Wireless | Wi-Fi 7 and Bluetooth |
| USB | 4 × USB Type-C |
| Display Output | 1 × HDMI 2.1a, plus DisplayPort over supported USB-C connections |
| Dimensions | 150 × 150 × 50.5 mm |
| Weight | Approximately 1.2kg |
Ideal For
- Local large language model inference and experimentation
- Generative AI and AI agent development
- AI model prototyping and fine-tuning within supported memory and compute limits
- Research, data science, and machine learning development
- Building and testing AI applications before deployment to data centres or cloud environments
What's in the Box
- NVIDIA DGX Spark system
- Power supply
- Quick-start documentation and support information
Warranty
Warranty coverage is subject to the authorised seller, distributor, and applicable regional terms. Please confirm the warranty duration at the time of purchase.
Note: Specifications are based on NVIDIA's published DGX Spark information. Confirm the exact storage configuration, included accessories, and regional warranty with the seller before ordering.
More Products
Before you buy
Is this genuine, warranted stock?+
Yes — authorised NVIDIA stock with full brand warranty, GST invoice, and serial tracking.
How fast does it ship?+
Dispatched in 2–7 days with tracked shipping across India. Tracking is shared by email and SMS.
What payment options are available?+
UPI, all major cards, net-banking, and no-cost EMI on eligible orders — chosen at checkout.
Can I return or replace it?+
Check our refund/return policies in the footer menu
I need technical help/assistance?+
Contact 8356026438 for any querries



