Axelera AI
Axelera AI is delivering the world’s most powerful and advanced solutions for AI at the Edge. Its industry-defining Metis™ AI platform – a complete hardware and software solution for AI inference at the edge – makes computer vision applications more accessible, powerful and user friendly than ever before. Based in the AI Innovation Center of the High Tech Campus in Eindhoven, The Netherlands, Axelera AI has R&D offices in Belgium, Switzerland, Italy and the UK, with over 195 employees in 18 countries. Its team of experts in AI software and hardware come from top AI firms and Fortune 500 companies.
- AI Processing Unit
- AI Accelerators: Card & Systems
- Voyager SDK
Metis AIPU: The Future of AI Acceleration at the Edge
What Is the Metis AIPU?
The Metis AI Processing Unit (AIPU) by Axelera AI is a groundbreaking AI accelerator designed specifically for edge computing applications. Unlike traditional processors that struggle to deliver high performance with low power consumption, Metis AIPU revolutionizes edge AI by combining extreme computational efficiency, scalability, and seamless deployment capabilities.
With its in-memory compute architecture, the Metis AIPU enables powerful, real-time AI inference directly on edge devices, eliminating the need for costly and time-consuming cloud interactions.
Key Features of Metis AIPU
Unmatched Performance Efficiency
- 214 TOPS (Tera Operations Per Second) in a single chip
- Industry-leading TOPS/Watt ratio, optimizing both performance and power
- Scalable across multiple devices for larger AI workloads
In-Memory Computing
- Reduces data movement by processing data where it is stored
- Minimizes latency and energy usage
- Ideal for computer vision and real-time inference tasks
Edge-Ready Design
- Compact, low-power footprint perfect for embedded systems
- Supports a wide range of edge AI applications: smart cameras, industrial automation, retail analytics, robotics, and more
Integrated Software Stack
The Metis AIPU is powered by Voyager SDK, a complete software suite that simplifies AI model deployment and optimization. It includes:
Compiler toolchain for converting popular AI models
Runtime environment optimized for edge devices
Support for TensorFlow, PyTorch, ONNX, and other major frameworks
This seamless integration accelerates time-to-market and lowers the barrier for developers transitioning AI workloads to the edge.
Use Cases: AI at the Edge, Redefined


Smart Surveillance
Detect and analyze events in real-time without sending data to the cloud.
Industrial Automation
Deploy AI models directly onto the factory floor sensors and controllers.


Retail Intelligence
Enable in-store analytics and shopper behavior tracking with instant processing.
Smart Cities & Mobility
Support advanced driver-assistance systems (ADAS), traffice monitoring, and more.
Why Choose Axelera's Metis AIPU?
Axelera AI’s Metis AIPU combines cutting-edge AI performance, ultra-low latency, and efficient power usage in a compact form factor. It is engineered to empower businesses and developers to harness the full potential of AI directly at the edge — where data is generated and decisions must be made instantly.
Discover Axelera AI’s Metis Product Line: Powering Edge AI with Unmatched Efficiency
As industries increasingly shift AI workloads from the cloud to the edge, the demand for high-performance, low-power AI hardware has never been greater. Axelera AI responds to this need with its Metis product line, a suite of AI acceleration platforms built around the Metis AIPU — a revolutionary in-memory computing architecture designed for real-time inference and computer vision at the edge.
Metis M.2 Card – Compact Power for Embedded AI

The Metis M.2 Card delivers full AI inference acceleration in an ultra-compact format, perfect for space-constrained and fanless edge devices.
Key Features:
Form Factor: M.2 2280 B+M key
Performance: 214 TOPS @ 8W
Interface: M.2 socket, standard in many embedded platforms
Compatibility: Linux-based embedded systems, smart cameras, drones, and robotics
Highlights:
Minimal thermal footprint for passive cooling
Direct integration into existing embedded designs
Ideal for AI on the move or in tightly enclosed environments
Metis PCIe Card – High Throughput for Edge Servers

The Metis PCIe AI Accelerator offers a high-performance, plug-and-play solution for AI inference in desktop, server, or industrial edge compute environments.
Key Features:
Form Factor: Half-Height, Half-Length (HHHL) PCIe Gen4 x8
Performance: 214 TOPS @ 15W
Scalability: Multi-card support for higher throughput
Compatibility: x86 edge servers, workstations, and industrial PCs
Highlights:
Exceptional TOPS/Watt performance
Ideal for smart cities, industrial vision, and real-time analytics
Enables AI compute in standardized hardware platforms
Metis Compute Board – Integrated Compute for Custom Systems

The Metis Compute Board is a complete, compact compute module integrating the Metis AIPU and designed for seamless integration into custom hardware designs.
Key Features:
Integrated AIPU and required I/O interfaces
Optimized for deep integration into custom edge devices
Suitable for OEMs and hardware manufacturers
Delivers production-ready compute in a small footprint
Highlights:
Offers full control and flexibility for embedded system designers
Ideal for mass production and custom hardware builds
Designed for applications like machine vision, industrial robotics, and smart infrastructure
Metis Evaluation Systems – Accelerate AI Prototyping
Axelera’s Metis Evaluation Systems provide a ready-to-use platform for developers to explore and benchmark the power of the Metis AIPU in real-world scenarios.
Includes:
Metis PCIe Card or Metis M.2 Card (depending on configuration)
Reference host system or integration board
Pre-installed Voyager SDK
Tools for model import, optimization, and deployment
Use Cases:
Proof of concept (PoC) development
Model benchmarking and testing
AI workload evaluation across edge use cases
Why it matters: The evaluation systems are the ideal starting point for R&D teams and innovators who need a full-stack environment to validate edge AI solutions before production.
Voyager SDK by Axelera AI: Simplifying Edge AI Development from Model to Deployment
As AI applications move increasingly to the edge, developers face the challenge of deploying powerful models on resource-constrained devices — without compromising accuracy or speed. The Voyager SDK from Axelera AI solves this challenge by offering a fully integrated software stack designed to accelerate and simplify the journey from AI model development to deployment on Metis AIPU-powered platforms.
Whether you're building smart cameras, autonomous machines, or industrial vision systems, Voyager SDK equips you with the tools to harness the full potential of in-memory computing — effortlessly.
Key Features of Voyager SDK
End-to-End Toolchain
Voyager SDK supports every stage of the AI lifecycle, including:
Model conversion from frameworks like TensorFlow, PyTorch, and ONNX
Model quantization and optimization for edge inference
Automated compilation for Metis AIPU hardware
Edge runtime environment for low-latency inference
From training to deployment, Voyager ensures smooth, efficient model delivery.
Optimized for In-Memory Compute
Voyager SDK is purpose-built to extract maximum performance from Axelera’s Metis AIPU, leveraging its in-memory compute architecture to:
Minimize data movement
Reduce energy consumption
Deliver ultra-low-latency inference
Achieve high throughput for vision workloads
This ensures your models run faster, leaner, and smarter on the edge.
Developer-Friendly Experience
Voyager SDK simplifies edge AI development through:
Command-line and API access for automation and scripting
Integrated profiling and debugging tools
Clear performance metrics and optimization insights
Support for containerized workflows (e.g., Docker)
It’s built to meet the needs of AI developers, ML engineers, and system integrators alike.
Flexible Model Input & Compatibility
Voyager offers wide compatibility with common frameworks and formats:
ONNX, TensorFlow, PyTorch
Pre-trained and custom models
Batch processing and streaming inference
Easy integration into CI/CD pipelines
Your existing model investments are preserved, with minimal effort needed to port them to Metis hardware.