A new desktop AI computer specifically designed for the complete construction and operation of AI. It utilizes the NVIDIA GB10 Grace Blackwell superchip, providing up to 1000 AI TOPS performance with energy efficiency and a compact design to power large AI workloads. With 128GB of unified address system memory, developers can proto- type, fine-tune, or infer next-generation AI models with up to 200 billion parameters and seamlessly deploy them to data centers or cloud platforms. Additionally, the integrated NVIDIA ConnectX™ network allows two AI computers to be connected, supporting inference on models with up to 405 billion parameters.
| Architecture | NVIDIA Grace Blackwell |
| CPU | 20-core Arm 10x Cortex-X925 + 10x Cortex-A725 |
| GPU | NVIDIA Grace Blackwell |
| Storage | 4 TB NVME.M2 with self-encryption |
| OS | NVIDIA DGX OS |
| System Memory | 128GB LPDDR5x Unified Memory |
| Price |
AED :
25,000.00
USD :
6,807.50
|
| Action | Get a Quote |
FusionXpark GB10, A new desktop AI computer specifically designed for the complete construction and operation of AI. It utilizes the NVIDIA GB10 Grace Blackwell superchip, providing up to 1000 AI TOPS performance with energy efficiency and a compact design to power large AI workloads. With 128GB of unified address system memory, developers can prototype, fine-tune, or infer next-generation AI models with up to 200 billion parameters and seamlessly deploy them to data centers or cloud platforms. Additionally, the integrated NVIDIA Connect™ network allows two AI computers to be connected, supporting inference on models with up to 405 billion parameters.
With its outstanding energy efficiency and compact design, this AI computer offers powerful AI supercomputing capabilities and superior performance, making it the ideal choice for AI developers, researchers, and data scientists.
Mainstream ecosystem of FusionXpark GB10 support, with one-stop workflow from model prototyping, fine-tuning, inference to AI agent deployment.
Ensuring data privacy and accelerating productivity, applicable to scenarios such as predictive maintenance, medical image analysis, and financial forecasting.
Effortless AI teaching platform setup, delivering "personal supercomputing" research experience with agile teaching and training.
End-to-end development of edge AI applications on desktop for real-time responsiveness and edge computing capability.
| 1 | Power button |
| 2 | Power port |
| 3 | USB-C (DisplayPort) |
| 4 | HDMI port |
| 5 | Ethernet port |
| 6 | QSFP |
| Architecture | NVIDIA Grace Blackwell |
| GPU | NVIDIA Grace Blackwell GB10 Superchip, up to 1 PetaFLOP AI (FP4), 6144 CUDA/Tensor/RT cores |
| CPU | 20-core Arm |
| CUDA Cores | 6,144 CUDA Cores |
| Tensor Cores | 5th Generation |
| RT Cores | 4th Generation |
| Tensor Performance | 1 PFLOPS |
| System Memory | 128GB LPDDR5x Unified Memory |
| Memory Interface | 256-bit |
| Memory Bandwidth | 273 GB/s |
| Storage | 4 TB NVME.M2 with self-encryption |
| USB | 4x USB TypeC |
| Ethernet | 1x RJ-45 connector 10 GbE |
| NIC | ConnectX-7 Smart NIC |
| Wi-Fi | WiFi 7 |
| Bluetooth | BT 5.4 |
| Audio-output | HDMI Multi-channel Audio Output |
| Display Connectors | 1 × HDMI 2.1a |
| NVENC | NVDEC | 1x | 1x |
| OS | NVIDIA DGX™ OS |
| System Dimensions | 150 mm L x 150 mm W x 50.5 mm H |
| System Weight | 1.2 kg |
| Warranty | 1 year warranty |
