ASUS Ascent GX10 Personal AI Supercomputer
Desktop AI Powerhouse Powered by NVIDIA GB10 Grace Blackwell Superchip
Bring petaFLOP-scale AI computing directly to your desk with the ASUS Ascent GX10. This compact personal AI supercomputer delivers up to 1,000 TOPS of AI performance in an ultra-small 1.6-liter form factor, making it the ideal solution for AI developers, researchers, and data scientists who demand local AI processing without recurring cloud computing costs.
Extreme AI Performance
Built on the NVIDIA GB10 Grace Blackwell Superchip with fifth-generation Tensor Cores and FP4 support, the Ascent GX10 delivers 1 petaFLOP of AI computing power. With 128GB of unified LPDDR5x memory, you can fine-tune, prototype, and run inference on large language models up to 200 billion parameters—all locally on your desktop.
Scalable Architecture
Need even more power? Connect two GX10 systems via the integrated NVIDIA ConnectX-7 SmartNIC with NVLink-C2C technology to handle massive models like Llama 3.1 with 405 billion parameters. The built-in 10GbE LAN and 200GbE RDMA networking enable seamless cluster expansion.
Compact & Efficient Design
At just 150 x 150 x 51 mm and weighing only 1.48kg, the GX10 fits anywhere—labs, offices, or edge environments. The advanced dual-fan thermal system with 7-level control ensures sustained peak performance in this remarkably small footprint.
Key Specifications
| Processor | NVIDIA GB10 Grace Blackwell Superchip (ARM v9.2-A Cortex X925) |
| GPU | NVIDIA Blackwell GPU (integrated) |
| Memory | 128GB LPDDR5x Unified System Memory |
| Storage | 1TB M.2 NVMe PCIe 4.0 SSD |
| Networking | 10GbE LAN, NVIDIA ConnectX-7 SmartNIC, Wi-Fi 7, Bluetooth 5.4 |
| I/O Ports | 4x USB 3.2 Gen 2x2 Type-C (20Gbps), HDMI 2.1 |
| Operating System | Ubuntu Linux (DGX OS) |
| Power | 240W Adapter |
| Dimensions | 5.91 x 5.91 x 2.01 inches (150 x 150 x 51 mm) |
| Weight | 3.26 lbs (1.48kg) |
| Color | Stellar Grey |
Full-Stack AI Solution
Leverage NVIDIA's complete AI software stack for seamless development. Models built on the GX10 can be transitioned to NVIDIA DGX Cloud or any accelerated data center infrastructure with minimal code adjustments—perfect for prototyping before production deployment.
Ideal for: AI Development, Machine Learning Research, LLM Fine-Tuning, Data Science, Edge AI Deployment
