Home Hailo Hailo-8 M.2 AI Acceleration Module
Hailo

Hailo-8 M.2 AI Acceleration Module

The Hailo-8 M.2 AI Acceleration Module is a high-performance AI accelerator designed to bring 26 TOPS of processing power to edge devices. This M.2 form factor module supports PCIe Gen-3.0 with 2 or 4 lanes, offering real-time, low-latency AI inferencing with exceptional power efficiency. It is compatible with various edge devices, supporting multi-stream and multi-model processing. The module enables quick deployment of AI-powered edge applications in industries requiring high-efficiency deep learning models and rapid time-to-market..

  • Detail
  • Special Features
  • Specification

The Hailo-8 M.2 AI Acceleration Module integrates the Hailo-8 AI processor into an M.2 module form factor, designed to deliver 26 TOPS of AI processing power for edge devices. This module enables real-time deep learning applications with low-latency and high efficiency, making it perfect for industries that require high-performance AI capabilities in compact, power-efficient packages.

With support for multiple AI frameworks such as TensorFlow, TensorFlow Lite, ONNX, Keras, and PyTorch, the Hailo-8 M.2 module ensures seamless compatibility with a wide range of AI models. It features PCIe Gen-3.0 interfaces with options for 2 lanes (16 Gbs) or 4 lanes (32 Gbs), offering flexible and scalable integration into edge devices.

The module supports simultaneous multi-stream processing, allowing it to handle multiple AI tasks in parallel without compromising on performance. Additionally, it includes a comprehensive software suite, enabling easy integration and deployment of deep learning models with minimal effort.

The Hailo-8 M.2 module is ideal for applications that require high performance in compact devices, including video analytics, surveillance systems, industrial monitoring, and autonomous systems. The module also supports industrial-grade conditions, with an operating temperature range of -40°C to 85°C for reliable operation in various environments.

  • 26 TOPS of AI Performance

    • Powered by the Hailo-8 AI processor, the Hailo-8 M.2 module delivers 26 Tera-Operations per Second (TOPS), enabling real-time AI inferencing for edge applications.

  • Best-in-Class Power Efficiency

    • Designed for edge devices, the module offers best-in-class power efficiency, consuming minimal power while maximizing AI processing performance.

  • Real-Time, Low-Latency AI Inferencing

    • Low-latency processing for deep neural network models, making it ideal for applications in industrial automation, surveillance, smart cities, and more.

  • Cost-Efficiency

    • Highest cost-efficiency (TOPS/$) compared to competing solutions, delivering excellent performance per dollar in AI-powered edge products.

  • Scalable Performance

    • Simultaneous processing of multiple streams and multiple models, offering flexibility for complex AI tasks and deep learning applications.

  • Temperature Range Support

    • The Hailo-8 M.2 module supports an extended temperature range of -40°C to 85°C, ensuring reliable performance in harsh environments.

  • Rapid Time-to-Market

    • The module utilizes a standard M.2 form factor, supporting key M, key B+M, and key A+E connectors for easy integration into existing systems, reducing development time.

General Features

  • AI Processing Power: 26 TOPS

  • Power Efficiency: Best-in-class

  • Temperature Range: -40°C to 85°C

  • PCIe Interface:

    • Key M: 4 lanes (up to 32 Gbs)

    • Key B+M: 2 lanes (up to 16 Gbs)

    • Key A+E: Supports standard M.2 form factors

  • Power Consumption: Low power with high efficiency

Supported Frameworks

  • AI Frameworks: TensorFlow, TensorFlow Lite, Keras, PyTorch, ONNX

  • Host Architectures: X86, ARM

  • Operating Systems: Linux, Windows

Software and Integration

  • Dataflow Compiler: Enables easy and quick porting of neural network models

  • Robust Software Suite: Supports deep learning models and AI applications out-of-the-box

Physical Characteristics

  • Dimensions:

    • 22 x 42 mm with breakable extensions to 22 x 60 mm and 22 x 80 mm

    • 22 x 30 mm for Key B+M

  • Certification: CE, FCC