Raspberry Pi AI HAT+ 2

Raspberry Pi AI HAT+ 2

The Raspberry Pi AI HAT+ 2 is an expansion board designed for the Raspberry Pi 5, built around the Hailo-10H AI accelerator, bringing generative AI capabilities directly to the device.

Featuring 8GB of dedicated onboard RAM, this add-on enables local execution of advanced models such as Large Language Models (LLMs) and Vision-Language Models (VLMs), while freeing up the Raspberry Pi 5 to handle additional tasks. The AI HAT+ 2 delivers consistent, low-latency performance with hardware-accelerated AI at the edge, making it well suited for applications such as robotics, facility management, offline process control and secure data analysis.

The AI HAT+ 2 provides up to 40 TOPS (INT4) of AI inference performance, while offering computer vision capabilities comparable to the Raspberry Pi AI HAT+ 26 TOPS. Hailo supplies a selection of sample models, and users can also develop and train custom models tailored to their needs, including speech-to-text, translation, and visual scene analysis.

Like its predecessor, the AI HAT+, this version connects via the Raspberry Pi 5’s PCI Express interface. When running an up-to-date Raspberry Pi OS, the system automatically detects the onboard Hailo accelerator and makes it available for AI workloads. The built-in rpicam-apps included in Raspberry Pi OS also support the module natively, leveraging the accelerator for compatible post-processing tasks.

Software for supported generative AI models is available through Hailo’s official GitHub repository.
Note: Raspberry Pi 5 is not included (sold separately).

Raspberry Pi AI HAT+ 2 Features

• Hailo-10H AI accelerator with up to 40 TOPS (INT4) inference performance
• Computer vision performance comparable to the 26 TOPS AI HAT+
• Efficient execution of generative AI models thanks to 8GB onboard RAM
• Seamless integration with Raspberry Pi camera software stack
• Fully compliant with Raspberry Pi HAT+ specification
• Includes optional heatsink, 16mm stacking header, spacers, and screws for installation on Raspberry Pi 5 (even with Active Cooler installed)

Resources

• Raspberry Pi AI HAT+ 2 documentation
• Raspberry AI HAT+ 2 product brief
• AI HAT+ 2 Launch blog post