ASUS Ascent GX10 AI Supercomputer (NVIDIA GB10 Superchip)
1TB PCIe NVMe SSD
$5,899.00
Skip to product information
ASUS Ascent GX10 AI Supercomputer (NVIDIA GB10 Superchip)
$5,899.00
The ASUS Ascent GX10 puts the NVIDIA GB10 Grace Blackwell Superchip in a stackable, toolless desktop chassis. It offers DGX Spark-class local compute for ML engineers and small AI teams, with 128GB of unified memory. It sits on a desk alone or stacks with a second unit when a project outgrows one machine's memory.
Who it's for
- Fine-tuning, inference and prototyping with 70B-parameter models in FP4/INT4 quantization
- Running 30B models at higher precision with a long context window
- Agentic AI development with OpenClaw and NVIDIA NemoClaw
- Teams pooling two units' unified memory over ConnectX-7 200GbE
Key features
- NVIDIA GB10 Superchip: An Arm v9.2-A 20-core CPU (10x Cortex-X925 + 10x Cortex-A725) joined to an integrated Blackwell GPU with 5th-gen Tensor Cores and 4th-gen RT Cores over NVLink-C2C, which ASUS states gives roughly 5x the bandwidth of PCIe Gen 5, rated up to 1 petaFLOP (1000 AI TOPS) of FP4 performance.
- 128GB unified LPDDR5x at 273GB/s: One coherent memory pool at 8533MT/s shared by CPU and GPU, so large language models fit that would exceed the VRAM of most discrete workstation GPUs.
- 1TB, 2TB or 4TB NVMe storage: A user-upgradeable M.2 2242 SSD reached through the base grille.
- Dual-port NVIDIA ConnectX-7 (200GbE): Networking for clustering multiple GX10 units, plus a dedicated 10GbE RJ-45 port.
- Toolless, stackable chassis: The power switch is on the front rather than the rear panel, which helps when units are stacked or rack-mounted.
- Dual-fan cooling: Active cooling engineered to stay quiet under sustained AI workloads, with improved thermal headroom versus comparably sized systems according to ASUS.
- NVIDIA DGX OS: Ubuntu-based, with CUDA, PyTorch and TensorFlow pre-installed, and support for agentic AI frameworks including OpenClaw and NVIDIA NemoClaw.
- Connectivity: 3x USB-C 3.2 Gen2x2 (20Gbps) carrying DisplayPort 2.1, 1x USB-C with 180W PD 3.1 EPR power delivery, HDMI 2.1, Wi-Fi 7 and Bluetooth 5.4.
Specifications
| Spec | Detail |
|---|---|
| Superchip | NVIDIA GB10 Grace Blackwell |
| CPU | Arm v9.2-A, 20-core (10x Cortex-X925 + 10x Cortex-A725) |
| GPU | Integrated NVIDIA Blackwell, 5th Gen Tensor Cores, 4th Gen RT Cores |
| CPU-GPU interconnect | NVLink-C2C |
| AI performance | Up to 1 petaFLOP / 1000 AI TOPS (FP4, with sparsity) |
| Memory | 128GB LPDDR5x, 8533MT/s, 273GB/s bandwidth |
| Storage | 1TB / 2TB (PCIe Gen4) or 4TB (PCIe Gen5) NVMe SSD, user-upgradeable M.2 2242 |
| Networking | Dual-port NVIDIA ConnectX-7 (200GbE), 10 Gigabit Ethernet (RJ-45) |
| Wireless | Wi-Fi 7, Bluetooth 5.4 |
| Ports | 3x USB-C 3.2 Gen2x2 (20Gbps) with DisplayPort 2.1, 1x USB-C with 180W PD 3.1 EPR, 1x HDMI 2.1 |
| Operating system | NVIDIA DGX OS (Ubuntu-based), CUDA/PyTorch/TensorFlow pre-installed |
| Agentic AI support | OpenClaw, NVIDIA NemoClaw compatible |
| Power supply | 240W external USB-C adapter |
| Dimensions | 150 x 150 x 51mm |
| Weight | 1.48kg |
| Cooling | Dual-fan active cooling |
| Chassis | Toolless, stackable, front-mounted power switch |
Good to know
- The 1 petaFLOP / 1000 TOPS figure is a peak FP4 number with sparsity, so real dense-precision throughput on a given model will be lower.
- This is a fine-tuning, inference and prototyping platform, not a from-scratch foundation-model trainer. The 140W GB10 package is cooled by the dual-fan design.
- The standard DGX OS and CUDA stack means workloads developed here are portable to other DGX Spark-class systems or DGX Cloud.
Choose the GX10 if you want a GB10 system with a front power switch, an upgradeable M.2 slot and a stackable design. It runs Arm, so it suits teams working in the CUDA ecosystem.