Axelera Metis AI Accelerators – now available at Rutronik

In the rapidly evolving world of artificial intelligence, edge computing is becoming the cornerstone of real-time, energy-efficient, and scalable AI deployment. At the forefront of this transformation is Axelera AI, a European pioneer whose Metis platform is setting new standards for AI inference at the edge with industry-leading performance, usability, and efficiency at a fraction of the cost of existing solutions. Now available through Rutronik, Axelera’s Metis products offer a compelling alternative to traditional GPU-based solutions—delivering unmatched performance, efficiency, and ease of integration.

What is Axelera Metis?

The Metis platform is a holistic AI solution combining cutting-edge hardware and software. At its core is the Metis AIPU (AI Processing Unit), a quad-core chip built on a proprietary RISC-V architecture and Digital In-Memory Computing (D-IMC). This architecture enables up to 214 TOPS of AI inference performance at 15 TOPS/W energy efficiency, making it ideal for computer vision inference on the edge, with scalable technology to meet the future demands of AI acceleration workloads.

Complementing the hardware is the Voyager SDK, a comprehensive software stack that simplifies model deployment, optimization, and integration. Whether you’re working with TensorFlow, PyTorch, or ONNX, Voyager ensures developers can hit the ground running.

Metis® M.2 AI Inference Acceleration Card

Axelera® AI’s Metis M.2 card is perfect for constrained devices that require high-performance AI inference.

The M.2 AI accelerator from Axelera AI offers top-level performance and is the best choice for speeding up AI tasks. It uses one Metis AIPU and has 1 GB of dedicated DRAM memory, which helps reduce energy use and makes it easier to include in systems. This makes AI processing accessible for all kinds of applications. It greatly improves any vision processing system that uses an NGFF (Next Generation Form Factor) M.2 socket.

Order Metis® M.2 Card at R24 Order Metis® M.2 Card w/cooling at R24 Metis® M.2 Card Product Brief

Technical specifications

M.2 2280 M-key Form Factor
PCIe Gen3 x4 Interface
1 GB DRAM dedicated memory
Up to 214 TOPS
5-9 W Typical Application Power
Security Features (Secure Boot, Root of Trust)
Optional active cooling solution

Key features:

State-of-the-art AI inference for small footprint devices.
Inference on multiple cameras / support for many independent parallel neural networks.
Wide range of end-to-end AI pipelines and models available.
Voyager® SDK for evaluation and software integration.
Advanced quantization tools for uncompromised prediction accuracy.

Target applications:

Surveillance: easy deployment in shops, building sites, factories, and venues.
Robotics: high resolution advanced gesture recognition, pose estimation and human intention decoding.
Medical: improved accuracy of real time diagnostics, monitoring, imaging and analysis.
Agritech: crop analysis, smart irrigation/pest control, automated harvesting.
Drones: navigation, object detection and surveillance in lightweight, battery-powered UAV systems.

Metis® PCIe AI Inference Acceleration Card

Axelera® AI’s Metis PCIe card offers the highest performance inference acceleration on the market. It combines ease of use, power efficiency, and scalability.

For developers looking to boost performance, Axelera AI offers the Metis® PCIe accelerators. These cards use one Metis AIPU and can provide up to 214 TOPS, making them capable of handling the toughest vision tasks. There will be two memory configurations: 4 GB and 16 GB. The powerful performance is further supported by the Voyager® SDK software, which makes it easier and faster to deploy applications.

The PCIe accelerator will soon be available in another version with four Metis chips, providing up to 856 TOPS. There will be two memory configurations: 16 GB and 64 GB.

Order Metis® PCIe Card (4GB) at R24 Order Metis® PCIe Card (16GB) at R24 Metis® PCIe Card Product Brief

Technical specifications

PCIe single slot, HHHL Form Factor
PCIe Gen3 x4 Interface
4 or 16 GB DRAM dedicated memory
Up to 214 TOPS
8-15 W Typical Application Power
Security Features (Secure Boot, Root of Trust)
Active air cooling solution
Up to 3.200 FPS for ResNet-50

Key features:

Highest-performance AI-accelerator PCIe card in the market for edge AI applications.
Inference on multiple cameras / support for many independent parallel neural networks.
Wide range of end-to-end AI pipelines and models available.
Voyager® SDK for evaluation and software integration.
Advanced quantization tools for uncompromised prediction accuracy.

Target applications:

Security: faster detection of incidents like abandoned baggage, intrusion, fall.
Industry 4.0: defect detection, quality control, worker safety.
Retail: in-store customer behavior, stock monitoring and automated checkout systems.
Mobility: tolling, enforcement and parking management.
Logistics: improved operational efficiency and safety of logistic centers.