ASUS Ascent GX10 GX10-GG0003BN GB10 128 GB LPDDR5x-SDRAM 1 TB SSD Mini PC Grey

Ultra-Small AI Supercomputer
Powered by NVIDIA DGX™ Spark
The ASUS Ascent GX10, accelerated by the NVIDIA GB10 Grace Blackwell Superchip and the NVIDIA AI Software Stack, provides a complete solution for developing and deploying AI. Its compact design facilitates seamless integration and deployment, enabling powerful AI performance for innovators who demand excellence. With advanced AI tools and NVIDIA® ConnectX®-7, this small-scale server extends your AI capabilities and supports your unique solutions.

Design
Revolutionary AI Performance on Your Desktop
The groundbreaking ASUS Ascent GX10 AI Supercomputer, equipped with the cutting-edge NVIDIA® GB10 Grace Blackwell Superchip from the NVIDIA DGX Spark, brings petaFLOP-scale AI computing power directly to the desks of developers, AI researchers, and data scientists. This innovative device is designed to empower local AI development with its exceptional performance and advanced features.

Compact, Powerful, and Scalable

Unmatched AI Performance
Up to 1 petaFLOP AI performance with FP4
Delivers up to 1 petaFLOP of AI performance for large AI workloads.

128 GB LPDDR5x Coherent Unified System Memory
Enables model development, experimentation, and inference with ample memory capacity.

Cutting-Edge Architecture
NVIDIA GB10 Grace Blackwell Superchip:
At the heart of the ASUS Ascent GX10, this advanced chip features a robust Blackwell GPU with fifth-generation Tensor Cores and FP4 support.

Powerful 20-Core Arm CPU:
Enhances data preprocessing and orchestration, accelerating model fine-tuning and real-time inference.

NVLink™-C2C Technology:
Provides a coherent CPU+GPU memory model with five times the bandwidth of PCIe 5.0.

Handles Large-Parameter AI Models
Supports AI models with 200 billion parameters
Prototype, fine-tune, and infer the latest AI directly on your desktop.​

Integrated NVIDIA® ConnectX®-7 Networking Technology
Connect two ASUS Ascent GX10 systems to tackle even larger models like Llama 3.1 with 405 billion parameters.

AI Networking
Meets Next-Generation Connectivity: NVIDIA® ConnectX-7
The ASUS Ascent GX10 is equipped with NVIDIA® ConnectX-7, offering ultra-high-speed networking that enables rapid data transfer and low-latency communication in distributed AI workloads.

Ultra-Fast AI Data Throughput
Up to 200 Gb/s bandwidth facilitates lightning-fast data transfer between nodes – ideal for large, distributed AI workloads.

Secure, Smart Networking
Integrated hardware acceleration for TLS, IPsec, and MACsec ensures encrypted data transfer without CPU overhead.

Precision-Critical Performance
IEEE 1588v2 PTP support enables microsecond-level time synchronization for time-critical AI and edge computing applications.

AI Experience
Integrated AI Software Stack for Seamless Development

Pre-Installed AI for Instant Development
NVIDIA DGX OS (Ubuntu-based) – Optimized AI environment, ready out-of-the-box.
NVIDIA AI Software Stack – Pre-installed frameworks, SDKs, and tools for rapid deployment.

Optimized AI Tools & AI Frameworks
CUDA, PyTorch, TensorFlow, Jupyter – Optimized for AI model development & inference.
NVIDIA TensorRT – High-performance AI inference engine.
NVIDIA NIMs & Blueprints – Pre-built AI workflows & microservices.

Industry-Leading AI Model Support
DeepSeek R1 – Optimized AI inference with up to 70B parameters.
Llama 3.1 – Generative AI with up to 405B parameters (Dual-GX10).
Meta, Google Models – Broad compatibility with industry-leading AI frameworks.

Scalable
Engineered for Maximum Efficiency

Optimized cooling design for sustained AI performance under heavy workloads
Compact form factor enables high-density AI computing in a small footprint

Application
Local Development, Scalable Deployment

Seamless Transition to Cloud
Leverage NVIDIA AI platform software architecture to transfer models from desktop environments to DGX Cloud or any other accelerated cloud or data center infrastructure with minimal code adjustments.

Cost-Effective Experimentation Platform
Free up critical compute resources in clusters better suited for training and deploying production models.