Axelera AI develops hardware and software for AI inference acceleration, with a focus on running computer vision and generative AI workloads efficiently at the edge. Its technology foundation combines proprietary digital in-memory computing with a RISC-V dataflow architecture to improve performance-per-watt, cost efficiency, and deployment usability for embedded and edge systems.
Axelera AI commercializes its technology as a platform that spans silicon, accelerator cards, single-board compute, ready-to-use systems, and deployment software. The company’s portfolio is centered on the Metis family for edge AI inference and is expanding upward with Europa and Titania to address more demanding workloads in edge servers, enterprise infrastructure, and high-performance AI environments.
Offerings, Capabilities, and Integrations
Axelera AI’s offering combines inference silicon with a software stack designed to shorten the path from model selection to deployment. Its capabilities center on local AI execution, multi-stream vision processing, support for multimodal and generative AI workloads, and integration into standard embedded and server-class hardware through familiar form factors such as M.2, PCIe, and single-board compute.
The software environment supports model compilation, optimization, and pipeline deployment through a simplified workflow, while remaining compatible with mainstream AI development practices and common frameworks. Axelera AI also works through an ecosystem of OEMs, software developers, system builders, distributors, and integrators to help customers move from evaluation to production deployments across industrial, commercial, and infrastructure use cases.
Products and Services
- Metis® AIPU: Axelera AI’s core AI processing unit for edge inference, built around the company’s digital in-memory computing and RISC-V-based architecture for high-efficiency computer vision and local AI workloads.
- Europa® AIPU: A higher-performance AIPU designed for more demanding edge and server-class inference workloads, extending the Axelera AI platform to larger visual, language, sensor, and speech processing applications.
- Voyager® SDK: Axelera AI’s software development kit for compiling, optimizing, and deploying AI inference pipelines on its hardware using a streamlined developer workflow.
- Voyager® SDK Model Zoo: A catalog of pre-trained models and turnkey pipelines integrated with the Voyager® SDK, covering tasks such as classification, detection, segmentation, keypoint detection, and face recognition.
- M.2 AI Inference Acceleration card: A compact M.2 accelerator card powered by a single Metis AIPU for embedded and space-constrained edge AI deployments.
- M.2 MAX AI Inference Acceleration card: An enhanced M.2 card based on the Metis AIPU that adds more memory bandwidth and higher memory options for more compute-intensive edge inference, including LLM and VLM workloads.
- PCIe AI accelerator card: A PCIe-based inference accelerator offering local AI processing in standard host systems, available in configurations suited to PCs, workstations, industrial systems, and edge servers.
- Metis® Compute Board: A single-board computer with an RK3588 processor and integrated Metis acceleration for compact, ready-to-deploy edge AI systems handling multi-stream computer vision and multimodal workloads.
- Metis Systems: Ready-to-use starter and development systems for x86 and Arm environments that combine Metis acceleration hardware with preinstalled Voyager® SDK software for evaluation and prototyping.
- Titania®: An announced chiplet architecture intended to scale Axelera AI’s digital in-memory computing approach beyond edge devices into enterprise, data center, and high-performance AI infrastructure.
Target Customers
Axelera AI targets OEMs, system builders, software developers, integrators, and channel partners that need to add high-performance inference to edge devices and on-premises systems without relying on cloud execution. Its products are suited to organizations that want standard hardware form factors, lower power draw, and simpler deployment workflows for production AI.
The company’s end-market focus spans security and surveillance, industrial automation, retail, smart city infrastructure, robotics, healthcare, agriculture, space, and high-performance compute. Axelera AI is especially relevant for customers building vision-centric systems, multi-camera analytics, quality inspection, safety monitoring, intelligent transportation, and other applications where latency, privacy, bandwidth, and power efficiency matter.
Key People
- Fabrizio Del Maffeo: Chief Executive Officer
- Jonathan Ballon: Chairman
- Evangelos Eleftheriou: Chief Technology Officer
- Marta Ostroumoff: Chief Financial Officer
- Alexis Crowell: Chief Marketing Officer; General Manager of the Americas
- Meghna Singh: Chief People Officer
- Louis Mather: Vice President, Strategy
- Flavio Devidé: Vice President, Sales
- Manuel Botija: Vice President, Product Management
- Jonathan Ferguson: Vice President, Silicon Engineering
- Ioannis Koltsidas: Vice President, AI Software
- Eke Bijzitter: Vice President, Legal
Key Facts
- Headquarters: Eindhoven, North Brabant, Netherlands
- Employees: 201-500
- Annual Revenue: Approximately $70M-$71M
- Parent Company: None
- Subsidiaries: None
- Publicly Listed: Private
