- Acer Veriton GN100 Use NVIDIA GB10 SUPERCHIP with 128 GB memory and 4 TB storage
- The compact design of less than 1.5 kg still offers 1 PFLOPS of IA FP4 performance
- Two scale of linked systems to handle models of AI with up to 405 billion parameters
Acer has announced the Veriton GN100 AI AI workstation, a compact desk built in the GB10 Grace Blackwell Superchip in Nvidia.
It is designed to execute large models of AI locally, with the option of climbing workloads even more linking two together systems.
The Veriton GN100, revealed in IFA 2025 in Berlin, offers up to 1 AI FP4 performance pflops through a combination of 20 CPU nuclei based on ARM, CUDA nuclei and fifth -generation tensioner nuclei.
Up to 405 billion parameters
The device comes with 128 GB of unified LPDDR5X memory and 4 TB of Self -Gering NVME storage.
Despite its tiny 150 mm x 150 mm footprint, and a U-5 kg weight, the GN100 Veriton is intended to deliver server class computing power in a PC Mini PC shape factor.
The system includes the NVIDIA AI software battery and the DGX operating system, with support for commonly used developers such as Pytorch, Jupyter and Ollama.
Acer says that this will allow developers, researchers and students prototype, refine and test large language models directly in the system, reduce the dependence on the infrastructure of the remote cloud and help reduce operating costs.
With a NVIDIA Connectx-7 Smartnic installed, two units of Veriton GN100 can be linked to the workloads at the scale, which allows the processing of AI models with up to 405 billion parameters.
This will allow working with even larger sets and training experiments without requiring immediate access to the data center hardware.
Connectivity includes four USB 3.2 type C, HDMI 2.1B, Ethernet, Wi-Fi 7 and Bluetooth 5.1.
Security characteristics include a Kensington block and local AI models execution to help protect confidential data.
Acer has established prices for the Veriton GN100 to $ 3,999 in North America and € 3,999 in EMEA, with availability that varies according to the region.
Acer is only the last PC manufacturer to launch a Nvidia DGX Spark clone, following the steps of Asus, HP and Dell, and MSI.
