Ascent GX10
A mini supercomputer released by ASUS with a volume of 1.5 liters, equipped with the NVIDIA GB10 Grace Blackwell superchip, and supporting two-unit stacking to scale development and inference for models with up to 405B parameters.
Timeline (last 90 days)
The ASUS Ascent GX10 weighs 1.6 kilograms and has a volume of 1.5 liters.
The Ascent GX10 is equipped with the NVIDIA GB10 Grace Blackwell superchip.
The Ascent GX10 supports dual-unit stacking expansion, supporting development and inference of models with up to 405B parameters.
A single GX10's native memory capacity is usually insufficient to independently host a full 405B model; the claimed maximum support for 405B relies on the aggregated memory pool after dual-unit stacking, along with quantization and tensor parallelism techniques.
The Ascent GX10 comes preinstalled with the NVIDIA AI software stack, covering the entire workflow from prototyping to deployment, and supports mainstream frameworks such as PyTorch and TensorFlow.
The front of the Ascent GX10 retains only the power button, with all IO ports concentrated on the back.
The application scenarios for the GX10 listed by the uploader include event broadcasting, modeling and rendering, biomedicine, and intelligent manufacturing.
The Ascent GX10 is positioned between high-end workstations and cloud computing power: more private and lower-latency than the cloud, and more compact and easier to deploy than traditional workstations.
The video content comes from a single uploader's unboxing demonstration; the actual inference performance of the 405B model, the specific efficiency of dual-unit stacking, and the stability of long-duration operation still require more independent testing and verification.
All Facts (9)
The ASUS Ascent GX10 weighs 1.6 kilograms and has a volume of 1.5 liters.
50%UnverifiedThe Ascent GX10 is equipped with the NVIDIA GB10 Grace Blackwell superchip.
50%UnverifiedThe Ascent GX10 supports dual-unit stacking expansion, supporting development and inference of models with up to 405B parameters.
50%UnverifiedA single GX10's native memory capacity is usually insufficient to independently host a full 405B model; the claimed maximum support for 405B relies on the aggregated memory pool after dual-unit stacking, along with quantization and tensor parallelism techniques.
50%UnverifiedThe Ascent GX10 comes preinstalled with the NVIDIA AI software stack, covering the entire workflow from prototyping to deployment, and supports mainstream frameworks such as PyTorch and TensorFlow.
50%UnverifiedThe front of the Ascent GX10 retains only the power button, with all IO ports concentrated on the back.
50%UnverifiedThe application scenarios for the GX10 listed by the uploader include event broadcasting, modeling and rendering, biomedicine, and intelligent manufacturing.
50%UnverifiedThe Ascent GX10 is positioned between high-end workstations and cloud computing power: more private and lower-latency than the cloud, and more compact and easier to deploy than traditional workstations.
50%UnverifiedThe video content comes from a single uploader's unboxing demonstration; the actual inference performance of the 405B model, the specific efficiency of dual-unit stacking, and the stability of long-duration operation still require more independent testing and verification.
50%