SciTransfer
Axelera Europa · Project

High-performance AI chips and software for real-time processing at the edge

digitalTestedTRL 5

Imagine if your smart devices had a brain as powerful as a giant data center but used very little electricity. This project builds a special chip and a simple software tool to make that happen. It allows machines to understand speech and images instantly without needing to send data to the cloud.

By the numbers
15.52 B
Projected edge AI market size in 2025 (EUR)
The business problem

What needed solving

Current AI models for vision and language are too power-hungry and expensive to run locally on edge devices. This creates a bottleneck for real-time applications in smart cities and industry 4.0.

The solution

What was built

The Europa AI-Core architecture and a corresponding System-on-Chip (SoC). It includes a software platform with a no-code web composer for easier AI deployment.

Audience

Who needs this

Embedded system manufacturersEdge datacenter operatorsIndustrial automation engineersSmart city infrastructure providers
Business applications

Who can put this to work

Industrial Automation
enterprise
Target: Smart Factory Operator

If you are a smart factory operator dealing with slow response times in quality control — this project developed a hardware and software platform that accelerates computer vision at a fraction of the costs and power consumption of current solutions.

Retail
mid-size
Target: Automated Store Provider

If you are an automated store provider dealing with high energy bills for AI cameras — this project developed the Europa AI-Core that enables real-time decision-making at the edge with high energy efficiency.

Security
any
Target: Smart City Surveillance Firm

If you are a surveillance firm dealing with the need for local natural language processing — this project developed a scalable SoC that works from a single embedded device to edge datacenters.

Frequently asked

Quick answers

How does this affect the cost of AI deployment?

The solution is designed to accelerate the latest generation of neural network models at a fraction of the costs of available solutions.

Can this be scaled for different business sizes?

Yes, the hardware is highly scalable, ranging from a single device in an embedded system to multiple instances on cards in an edge datacenter.

What is the IP or licensing model for the hardware?

Based on available project data, the project focuses on the development of the Europa AI-Core architecture and SoC, but specific licensing terms are not provided.

How easy is it to integrate into existing workflows?

The project includes a no-code web composer and a consistent development environment to drive the adoption of AI.

What is the timeline for the project deliverables?

The project period runs from 2024-03-01 to 2026-02-28.

Consortium

Who built it

The project is led by a single SME, Axelera AI BV from the Netherlands. This 100% industry-led consortium indicates a strong commercial drive, leveraging a leadership team with 20+ years of experience in major tech corporations like Asus and Advantech.

How to reach the team

Contact AXELERA AI BV in the Netherlands

Next steps

Talk to the team behind this work.

Contact us to explore integration opportunities with the Europa AI-Core.