- Ultra-compact personal AI workstation featuring the NVIDIA® GB10 Grace Blackwell Superchip, offering 128 GB unified memory, 4 TB storage, and up to 1 PFLOPS of FP4 AI performance.
- Designed to operate large AI models locally, with the ability to scale workloads by linking two Acer Veriton GN100 systems using NVIDIA ConnectX-7 SmartNIC, preloaded with the latest NVIDIA AI software stack and supporting popular developer tools such as PyTorch, Jupyter, and Ollama.
- Compact, secure, and versatile, with Kensington lock support and the capability to pair two devices to process AI models containing as many as 405 billion parameters.
Acer has announced the Veriton GN100 AI Mini Workstation, a compact yet highly capable system engineered to run large AI models locally. The device is designed to minimise reliance on cloud services while helping to cut related operational costs.
The workstation is powered by the NVIDIA® GB10 Grace Blackwell Superchip, providing up to 1 PFLOPS of FP4 AI performance. Combining advanced CUDA® cores, fifth-generation Tensor Cores, and 20 Arm-based CPU cores, the system is supported by 128 GB of unified system memory and 4 TB of NVMe M.2 SSD storage. This specification delivers server-grade performance within a mini-PC footprint.
Preinstalled with the NVIDIA AI software stack, the Acer Veriton GN100 delivers a complete platform for AI developers. Researchers, engineers, data scientists, and students can access widely used frameworks and tools such as PyTorch, Jupyter, and Ollama, enabling them to prototype, fine-tune, evaluate, and deploy large language models on-site, or expand seamlessly to cloud or data centre environments.
Through the NVIDIA ConnectX-7 NIC, two Veriton GN100 units can be connected to handle AI models of up to 405 billion parameters. The workstation also offers extensive connectivity and security features, including Wi-Fi 7, four USB 3.2 Type-C ports, an HDMI interface, Ethernet, and a Kensington lock, ensuring both reliability and secure deployment in diverse settings.