AI supercomputer, up to 1 petaFLOP performance, NVIDIA GB10 Grace Blackwell Superchip, NVIDIA DGX Spark, 128GB RAM, 1000GB SSD, OpenClaw and Hermes Agent support.

Compact desktop AI supercomputer powered by NVIDIA® GB10 Grace Blackwell Superchip and NVIDIA DGX™ Spark. Seamless integration and deployment, revolutionizing AI development and research.

- Extreme AI performance: The NVIDIA® GB10 Grace Blackwell Superchip provides 1 petaFLOP of AI performance and 128GB of memory to fine-tune the 200B model.
- Tailored platform for developers: Made specifically for AI researchers and developers, with full compatibility and simplified workflows.
- Scalable architecture: NVIDIA® NVLink™-C2C technology provides ultra-fast CPU-GPU memory communication, while the NVIDIA® ConnectX-7 network solution supports the connection of dual GX10 systems, thus guaranteeing outstanding scalability and performance.
- Advanced cooling design: The designed cooling ensures permanently high performance and reliability in an ultra-small design.
- Minimum footprint, maximum efficiency: High-density computing power in a compact system designed for laboratories, offices and edge environments.
- Complete AI solution: GB10 and the NVIDIA AI software package provide a complete solution for the development and implementation of artificial intelligence.

NVIDIA DGX™ Spark-based, ultra-small AI supercomputer

ASUS Ascent GX10, powered by the NVIDIA GB10 Grace Blackwell superchip and the NVIDIA AI software suite, provides a comprehensive end-to-end solution for all phases of AI development and deployment. Its compact design facilitates seamless integration and deployment, and delivers powerful AI performance for innovators who demand excellence. With advanced AI tools and NVIDIA®ConnectX®-7, this small server enhances AI capabilities while enabling custom solutions.

Design: Revolutionary AI performance on your desktop

The groundbreaking ASUS Ascent GX10 AI supercomputer, powered by the state-of-the-art NVIDIA®GB10 Grace Blackwell Superchip in NVIDIA DGX Spark, brings petaflop-scale AI computing capabilities directly to the desktop of developers, AI researchers and data scientists. This innovative tool is designed to enable local AI development with exceptional performance and advanced features.

Compact, powerful and scalable

- Compact size: 150 x 150 x 51mm

Unparalleled AI performance

- Provides up to 1 petaFLOP of AI performance, suitable for running large-scale AI workloads.
- 128 GB LPDDR5x coherent, unified system memory, which provides ample memory capacity for model development, testing and running (inference).

State-of-the-art architecture

- NVIDIA GB10 Grace Blackwell Superchip: At the heart of the ASUS Ascent GX10 is this advanced chip, which includes a high-performance Blackwell GPU with fifth-generation Tensor cores and FP4 support.
- High-performance 20-core Arm CPU: Accelerates data preparation and control, facilitating model fine-tuning and real-time inference.
- NVLink™-C2C technology: Provides a unified CPU+GPU memory model that offers five times the bandwidth compared to PCIe 5.0.

Efficiently handles large-scale generative AI models

- Support for AI models with 200 billion parameters
- Prototyping, fine-tuning and inferring the latest AI models right on your desktop.
- Integrated NVIDIA® ConnectX®-7 network technology
- By connecting two ASUS Ascent GX10 systems, even larger models such as Llama 3.1 with 405 billion parameters can be handled.

AI Network: Supports next-generation connectivity: NVIDIA® ConnectX-7

ASUS Ascent GX10 provides an integrated NVIDIA®ConnectX-7 ultragy network that enables fast data transfer and low-latency communication in distributed AI workloads.

Ultra-fast AI data transfer

Bandwidth of up to 200Gb/s enables lightning-fast data transfer between nodes - ideal for large-scale, distributed AI workloads.

A secure, intelligent network

It supports TLS, IPsec and MACsec encryption with hardware acceleration, thus ensuring secure data transfer without CPU load.

Precision-critical performance

IEEE 1588v2 PTP support enables microsecond-level time synchronization for time-sensitive AI and edge computing applications.

AI experience: Integrated AI Software Stack for seamless development

Pre-installed AI for instant development

- NVIDIA DGX OS (based on Ubuntu) Optimized AI environment, ready to use.
- NVIDIA AI Software Stack Pre-installed frameworks, SDKs and tools for quick deployment.

Optimized AI tools & AI frameworks

- CUDA, PyTorch, TensorFlow, Jupyter Optimized for artificial intelligence model development