.svg)
NVIDIA DGX Spark? belongs to a new class of computers designed to build and run AI, enabling developers to prototype, fine-tune, and inference large AI models locally and easily deploy to the data center or cloud. It is described as a “pint-sized AI supercomputer” that delivers up to 1 petaFLOP of AI performance in a compact form factor.
Specifications
| Feature | Details |
| Architecture | NVIDIA Grace Blackwell |
| GPU | Blackwell Architecture |
| CPU | 20-core Arm (10 Cortex-X925 + 10 Cortex-A725) |
| Tensor Cores | 5th Generation |
| RT Cores | 4th Generation |
| Tensor Performance | Up to 1 PFLOP (FP4) |
| System Memory | 128GB LPDDR5x, coherent unified system memory |
| Memory Interface | 256-bit |
| Memory Bandwidth | 273 GB/s |
| Storage | 1TB or 4TB NVMe M.2 with self-encryption |
| Operating System | NVIDIA DGX? OS |
Connectivity & Physical Specs
| Feature | Details |
| Ethernet | 1x 10 GbE RJ-45 connector |
| NIC | ConnectX-7 Smart NIC @ 200 Gbps |
| Wireless | Wi-Fi 7 and Bluetooth 5.4 |
| USB | 4x USB Type-C |
| Display Connectors | 1x HDMI 2.1a |
| Audio Output | HDMI multichannel audio output |
| Video Engines | 1x NVENC, 1x NVDEC |
| Power Consumption | 240W |
| System Weight | 1.2 kg |
| Dimensions | 150 mm L x 150 mm W x 50.5 mm H |
