Gigabyte AI TOP ATOM GB10 WiFi 7 LPDDR5x Integrated SoC 1L Motherboard

The Gigabyte AI TOP ATOM AI Workstation is a specialised computing node designed for on-premises machine-learning workloads. It handles tasks such as model training, fine-tuning, and inference, allowing researchers to process data without relying on external cloud services. At its core is a fully integrated NVIDIA GB10 Grace Blackwell Superchip, which combines processing cores and a unified architecture into a single SoC. The integrated board design minimises latency compared with modular desktop components, making it suited to developers working in labs or workshops who need dedicated hardware in a compact footprint.

Data retrieval and network connectivity shape the physical layout of this dedicated machine. As an on-site AI computing node, its external interfaces are configured to receive large training datasets quickly. Storage relies on a direct-attached solid-state drive, ensuring that large machine-learning models load directly into the processing pipeline. The system supports local monitoring via a display output when screen access is needed. Power comes through a standard AC cable, avoiding the complex server power supplies that can complicate deployment.

• 128 GB LPDDR5x system memory at 8533 MHz for high-speed model data access
• 1 TB M.2 NVMe PCIe 4.0 SSD providing local storage for fast dataset loading
• 10GbE RJ45 Ethernet interface delivering stable, low-latency wired networking for intensive data transfers
• Wi-Fi 7 (802.11be) support enabling high-throughput wireless operation in modern research environments
• Bluetooth 5.4 connectivity for direct integration with wireless peripherals and input devices
• HDMI 2.1a video port for direct high-resolution monitor output
• 1-litre chassis form factor minimising the physical footprint for space-constrained professional setups