ASUS has introduced the Ascent GX10, a compact AI supercomputer powered by NVIDIA’s new GB10 Grace Blackwell Superchip, bringing petaflop-scale computing performance to the desktop. Engineered for AI developers, researchers, and data scientists, the GX10 delivers up to 1,000 AI TOPS and supports 128GB of unified system memory, enabling local experimentation and inference of large generative AI models with up to 200 billion parameters.
At the heart of the GX10 is the NVIDIA GB10 Superchip, combining a Blackwell GPU featuring fifth-generation Tensor Cores with FP4 precision and a 20-core Arm Grace CPU. The superchip leverages NVLink-C2C technology to provide a unified CPU-GPU memory model with 5x the bandwidth of PCIe 5.0. Developers can prototype and fine-tune reasoning AI models—such as Llama 3.1 with over 400 billion parameters—by linking two GX10 systems via integrated NVIDIA ConnectX-7 NICs, making it an ideal platform for both local and scalable private-cloud deployments.
ASUS positions the Ascent GX10 as a flexible solution that bridges desktop AI development with large-scale cloud environments, including seamless migration to NVIDIA DGX Cloud and accelerated data center infrastructure. The system is designed to offload local development workloads, helping organizations reduce pressure on high-performance clusters reserved for full production models.
Key points:
• ASUS launches Ascent GX10, a desktop AI supercomputer delivering up to 1,000 TOPS of AI performance.
• Powered by the NVIDIA GB10 Grace Blackwell Superchip with integrated Blackwell GPU and Grace CPU.
• Supports 128GB unified memory and manages AI models with up to 200 billion parameters.
• Features NVLink-C2C for unified CPU+GPU memory bandwidth at 5x PCIe 5.0 speed.
• Scalable via ConnectX-7 NICs for larger workloads across two GX10 units.
• Seamless integration with NVIDIA AI software for easy transition to DGX Cloud and other infrastructure.
“AI is transforming every industry, and the ASUS Ascent GX10 is designed to bring this transformative power to every developer’s fingertips,” said KuoWei Chao, General Manager of ASUS IoT and NUC Business Group.
