Mini PC Workstation Acer Veriton VGN100 GB10, 128 GB LPDDR5x, 4 TB SSD, NVIDIA Blackwell, DGX OS, Black/Grey

The Acer Veriton GN100 AI Mini Workstation is a compact on‑premises AI-ready mini PC built around the NVIDIA GB10 (Blackwell) Superchip and DGX OS. Designed to run large AI models locally, it provides up to 1 petaFLOP of AI compute, 128 GB of unified LPDDR5x memory and up to 4 TB NVMe SSD storage in a space‑efficient chassis. This mini workstation supports high‑speed networking, stackable deployment and on‑device model development to reduce reliance on cloud resources and maintain data and model privacy.

Advantages

The device delivers server‑class AI performance in a desktop‑friendly mini PC form factor, combining a GB10 processor with 128 GB LPDDR5x and a 4 TB SSD to handle large datasets and demanding models. Built with the NVIDIA AI software stack and DGX OS, it enables developers and researchers to prototype, fine‑tune and run LLMs locally while offering scalable network options (ConnectX‑7) for multi‑unit setups. Security features such as a Kensington lock slot and a compact, stackable design support secure, space‑efficient deployments.

Suitable for

• Developers building and testing AI models locally (PyTorch, Jupyter, Ollama and similar toolchains)
• Universities and schools that need preconfigured AI workstations for teaching and research
• Data scientists analyzing large datasets with unified memory and fast storage
• Researchers running simulations and high‑performance AI experiments
• Startups and labs requiring on‑premises model privacy and reduced cloud costs

Key specifications

• Processor model: NVIDIA GB10 (Blackwell) with up to 1 petaFLOP AI performance
• Memory: 128 GB LPDDR5x‑SDRAM (8 x 16 GB unified memory)
• Storage: up to 4 TB NVMe M.2 SSD
• Operating system: DGX OS, 64‑bit
• Networking: Ethernet (10/100/1000/10000 Mbit/s), Wi‑Fi 7 (802.11be), Bluetooth 5.4
• Power supply: 240 W
• Chassis: Mini PC, colours: Black / Grey; weight: 1.2 kg

How to use

Install the Acer Veriton GN100 on a stable surface or stack multiple units as needed, secure the chassis using the Kensington lock slot and connect power and network. Use DGX OS to configure the system, then install or run AI frameworks such as PyTorch, Jupyter and other NVIDIA‑compatible tools to prototype and train models locally. For larger models or higher capacity, link units via high‑performance ConnectX‑7 networking to scale to multi‑unit clusters.

Recommendations

For optimal performance, keep DGX OS and NVIDIA drivers up to date, ensure adequate cooling and use NVMe storage for datasets and model checkpoints. When scaling across units, verify network configuration for ConnectX‑7 and monitor unified memory usage to balance workloads.