Cutting-edge research for the European AI market

The NeAIxt research initiative tackles a pressing question in contemporary AI and semiconductor development: How can Europe reduce its dependence on foreign technologies for edge artificial intelligence while delivering competitive performance, security, and efficiency? This EU-funded consortium aims to build a European alternative by advancing hardware components tailored to edge AI applications that outperform or complement existing solutions.

Strategic Objective: Strengthening Europe’s Competitiveness in Edge AI
Edge AI denotes deploying artificial intelligence directly where data is generated—on devices at the network’s edge—instead of relying on cloud-based computation. This approach reduces latency, enhances data privacy, and ensures operation under intermittent connectivity. European industry and research institutions see edge AI as critical for sectors ranging from automotive systems to industrial automation. However, achieving competitive edge AI capabilities requires innovations in both processing hardware and memory technology.

The NeAIxt project, launched in late 2025 and running through 2028, brings together more than 60 partners from across the EU to address this challenge. The consortium’s ambition is to boost European autonomy and influence in edge AI by developing state-of-the-art microcontrollers and storage technologies that meet modern demands for performance, energy efficiency, and security.

Technical Focus: Non-Volatile Memory and AI-Capable Microcontrollers
A central pillar of NeAIxt is the integration of novel embedded non-volatile memory (eNVM) into microelectronic systems. Traditional volatile memories require constant power to retain data, while non-volatile alternatives maintain data without power, enabling energy savings and resilience in edge devices. Fraunhofer IPMS, a core partner in the consortium, is advancing ferroelectric materials, such as optimized HfO₂ layers, that are capable of storing electrical polarization states permanently. Such layers are being embedded into established CMOS technologies to create efficient eNVM elements.

Alongside eNVM, the project focuses on designing microcontrollers with built-in AI processing capabilities. These microcontrollers aim to balance high computational throughput with low power consumption, a combination essential for real-time, on-device AI tasks without compromising system safety or responsiveness. Integrating AI accelerators directly within MCU architectures positions the NeAIxt hardware for applications where energy budgets and security constraints are stringent, a differentiation from many conventional edge AI platforms that rely on external accelerators or cloud resources.

From Research to Demonstrators
To validate the technical concepts, NeAIxt plans to deliver at least two demonstrators by the project’s completion. One demonstrator will showcase non-volatile data storage based on ferroelectric HfO₂ memory, and another will illustrate a hardware-based AI accelerator operating within the edge environment. Each demonstrator will undergo electrical characterization and benchmarking against existing technologies to assess gains in efficiency and system robustness.

These demonstrators serve multiple purposes: they provide empirical evidence of performance advantages, support future commercialization, and offer a testbed for optimizing integration into real-world systems. By establishing working prototypes that combine eNVM and AI processing within secure microcontrollers, NeAIxt aims to reduce barriers to adoption and build confidence in European-developed edge AI components.

Broader Impact: Energy Efficiency and Technological Sovereignty
The research undertaken in NeAIxt is motivated not only by competitive positioning but also by the broader need for energy-efficient computing infrastructure. Without advances in low-power hardware, the rapidly growing deployment of AI systems could significantly increase global energy consumption. By emphasizing energy efficiency and system security, NeAIxt aligns with EU objectives to sustain innovation while controlling environmental and strategic risks.

In summary, the NeAIxt project represents a concerted effort to develop European-centric solutions for edge AI that integrate cutting-edge memory technology with AI-ready microcontrollers. By producing validated demonstrators and fostering collaboration across research and industry partners, the initiative seeks to narrow technological gaps, enable secure and efficient AI at the edge, and strengthen Europe’s competitive stance in the global semiconductor and AI markets.

www.ipms.fraunhofer.de