Hailo SC1438 AI Kit for Raspberry Pi 5: M.2 HAT+ and Hailo-8L 13 TOPS module
67,80 € INCL. VAT
- M.2 HAT+ supplied with a pre-installed Hailo-8L module
- 13 TOPS inference accelerator, PCIe 2.0 link
- For Raspberry Pi 5, integrated into the camera software stack
Business, school or public body? business account Education offers
- Description
- Additional information
Description
The Hailo AI Kit (Raspberry Pi AI Kit, ref. SC1438) adds a neural network inference accelerator to a Raspberry Pi 5. It combines an M.2 HAT+ board and a Hailo-8L AI acceleration module in M.2 2242 format, already installed on the board, which delivers 13 TOPS (13 trillion operations per second).
The M.2 HAT+ connects the module's M.2 connector to the PCIe 2.0 interface of the Raspberry Pi 5. A thermal pad fitted at the factory between the module and the board spreads heat across the components. The whole assembly complies with the Raspberry Pi HAT+ specification.
On the software side, the kit is integrated into the Raspberry Pi camera software stack. To choose a model, Raspberry Pi points to Hailo's Model Explorer and describes how to run vision models in its dedicated AI software documentation.
Highlights
- 13 TOPS: acceleration module based on the Hailo-8L chip.
- Already assembled: the module and its thermal pad are pre-installed on the M.2 HAT+.
- PCIe 2.0 link: connection to the PCIe port of the Raspberry Pi 5 via ribbon cable.
- Stackable: 16 mm stacking GPIO header included.
- Software integration: supported in the Raspberry Pi camera software stack.
Technical specifications
- Manufacturer reference: SC1438
- Chip: Hailo-8L
- Stated performance: 13 TOPS
- Module format: M.2 2242, M-key connector
- Interface to the Raspberry Pi: PCIe 2.0
- Carrier board: M.2 HAT+, compliant with the Raspberry Pi HAT+ specification
- Compatibility : Raspberry Pi 5 only, with up-to-date Raspberry Pi OS
Use cases
- Computer vision: run vision models locally on the stream from a Raspberry Pi camera.
- Prototyping: add accelerated inference to an existing Raspberry Pi 5 project.
- Learning: discover how to run neural networks on a dedicated accelerator.
Compatibility and what you need to know
The kit is designed for the Raspberry Pi 5 only, as it uses its PCIe port. To mount it, you will also need a Phillips screwdriver. Raspberry Pi recommends adding an optional Active Cooler, and always disconnecting the power before working on the M.2 slot.
Do not confuse this kit with the AI HAT+ 13 TOPS (ref. SC1430). These are two separate products: the AI Kit is an M.2 HAT+ fitted with a Hailo-8L module in M.2 format, whereas the AI HAT+ integrates its accelerator directly on the board.
Raspberry Pi no longer manufactures the AI Kit and now directs new buyers to the AI HAT+.
Box contents
- 1 M.2 HAT+ board with pre-installed Hailo AI acceleration module and thermal pad in place
- 1 mounting kit: ribbon cable, spacers and screws
- 1 stacking GPIO header, 16 mm
Frequently asked questions
Is the Raspberry Pi 5 included?
No, the kit mounts on a Raspberry Pi 5 that you already own.
Does it work on a Raspberry Pi 4?
No, Raspberry Pi intends it only for the Raspberry Pi 5.
What is the difference from the AI HAT+ 13 TOPS?
The stated performance level is the same, 13 TOPS. The AI Kit uses an M.2 module mounted on an M.2 HAT+, whereas the AI HAT+ integrates the accelerator on its board.
Is cooling necessary?
Raspberry Pi describes the Active Cooler as optional but recommended for this setup.
Additional information
| HAT type | Artificial intelligence |
|---|
