Join the 155,000+ IMP followers

electronics-journal.com

IBASE Debuts PI900 Pico-ITX Motherboard for Edge AI

The ultra-compact platform integrates Intel Core Ultra processors and high-capacity DDR5 memory for space-constrained applications.

  www.ibase.com.tw
IBASE Debuts PI900 Pico-ITX Motherboard for Edge AI

IBASE Technology Inc. has introduced the PI900, an ultra-compact Pico-ITX motherboard engineered to deliver integrated AI acceleration and high-speed data processing at the edge. Powered by Intel® Core™ Ultra 100U Series processors—which combine CPU, GPU, and NPU capabilities (up to 12 cores and 14 threads)—the board supports an impressive 48GB of DDR5-5600 memory. Measuring just 100mm by 72mm, this highly integrated platform is specifically designed to handle data-intensive workloads in space-constrained environments, such as smart retail kiosks, machine vision systems, and other demanding embedded applications.

To ensure maximum configuration flexibility for system integrators, the PI900 features versatile display outputs, including HDMI 2.0, DisplayPort 1.4a via USB Type-C, and LVDS. Its robust expansion architecture includes dual M.2 sockets (M-Key for NVMe/SATA storage and E-Key for WiFi/Bluetooth), alongside a 2.5GbE LAN port, multiple USB interfaces, and RS-232/422/485 serial ports. Enhanced with a discrete Trusted Platform Module (TPM), hardware monitoring, and a 256-level watchdog timer, the PI900 provides the security, scalability, and reliability required for continuous industrial operation.

Additional Context
This section provides technological and market background not explicitly detailed in the original release.

As artificial intelligence moves from centralized cloud servers to the network edge, embedded hardware must balance intense computational demands with strict thermal and spatial limitations. The Pico-ITX form factor, which is roughly the size of a standard credit card, is traditionally utilized for basic IoT gateways or digital signage. However, the integration of Intel's Core Ultra processors introduces a dedicated Neural Processing Unit (NPU) directly onto the silicon. This hardware-level AI acceleration allows the board to efficiently execute machine learning inference tasks—such as real-time object detection in retail or defect analysis in factory machine vision—without the need to offload data to a discrete, power-hungry GPU. Paired with high-bandwidth DDR5 memory, this architecture significantly reduces latency and power consumption in fanless or ultra-compact industrial deployments where physical space and thermal dissipation are at an absolute premium.

Edited by Lekshman Ramdas, Induportals editor – adapted by AI.

www.ibase.com

  Ask For More Information…

LinkedIn
Pinterest

Join the 155,000+ IMP followers