- Memory
- 128 GB LPDDR5x unified
- Memory bandwidth
- 273 GB/s
- AI performance
- Up to 1 PFLOP FP4
- Storage
- 4 TB NVMe M.2, self-encrypting
NVIDIA DGX Spark 128GB / 4TB
P/N940-54242-0006-000
In transitETA confirmed on quote
Request a quote now
Price and availability confirmed within one business day
- Manufacturer channel supply
- Ships from Dubai / Hong Kong
- Commercial invoice, packing list, warranty docs
01Overview
DGX Spark is a compact desktop AI computer built on the NVIDIA GB10 Grace Blackwell Superchip with 128 GB of memory shared between CPU and GPU. It is bought by developers and small teams who want to prototype, fine-tune and run models of up to ~200B parameters locally with the same NVIDIA AI software stack as a data center DGX, then move the work to the cloud or a cluster.
Built for
- Local prototyping and inference of LLMs up to ~200B parameters (FP4)
- Fine-tuning of small and mid-size models on the desk
- Development against the full NVIDIA AI stack (DGX OS, CUDA, NIM) before deploying to DGX Cloud or clusters
- Pairing two units over ConnectX-7 for larger models
02Specifications
Technical specifications
Compute
- Superchip
- NVIDIA GB10 Grace Blackwell
- GPU
- Blackwell architecture, 6,144 CUDA cores, 5th-gen Tensor Cores, 4th-gen RT Cores
- CPU
- 20-core Arm (10× Cortex-X925 + 10× Cortex-A725)
- Tensor performance
- Up to 1 PFLOP FP4with sparsity
Memory
- System memory
- 128 GB LPDDR5x, coherent unified CPU/GPU memory
- Interface
- 256-bit, 4266 MHz
- Bandwidth
- 273 GB/s
- Storage
- 4 TB NVMe M.2 with self-encryptionA 1 TB variant also exists — this catalog item is the 4 TB configuration
Interface & form factor
- Networking
- NVIDIA ConnectX-7 NIC, 200 Gbps (2× QSFP); 1× 10 GbE RJ-45; Wi-Fi 7; Bluetooth 5.4
- USB
- 4× USB Type-C
- Display
- 1× HDMI 2.1a; up to 3× DisplayPort over USB-C
- Media engines
- 1× NVENC, 1× NVDEC
- Dimensions
- 150 × 150 × 50.5 mm
- Weight
- 1.2 kg
Power & cooling
- Power supply
- 240 W external adapter (included)
- GB10 TDP
- 140 W
Software & features
- Operating system
- NVIDIA DGX OS
- Software stack
- NVIDIA AI software stack (CUDA, NIM, frameworks) preinstalled
03Requirements
Installation
- Desk space of 150 × 150 mm and a mains outlet for the 240 W adapter — no rack or special cooling needed
- Monitor, keyboard and mouse connect over HDMI 2.1a / USB-C
- For two-unit clustering, QSFP cables or transceivers for the ConnectX-7 ports are ordered separately
- Confirm storage configuration (1 TB vs 4 TB) on the order
In the box
- DGX Spark system
- 240 W external power supply
04Documents
Supplied with every shipment
- Commercial invoice and packing list
- Manufacturer warranty registration documents
- Certificate of origin on request
- Serial numbers listed on the delivery note