The fundamental goal of the ANDANTE project is to leverage innovative hardware platforms to build strong hardware / software platforms for artificial neural networks (ANN) and spiking neural networks (SNN) as a basis for future products in the Edge IoT domain, combining extreme power efficiency with robust neuromorphic computing capabilities and demonstrate them in key application areas. The main objective of ANDANTE is to build and expand the European eco-system around the definition, development, production and application of neuromorphic hardware through an efficient cross-fertilization between major European foundries, chip design, system houses, application companies and research partners, as presented by the European Leader Group (ELG). The project brings together world class expertise to bring the world class expertise and infrastructures of Imec, CEA and FhG together with semiconductor companies, fabless, system houses, SMEs and application experts to explore and demonstrate the capabilities provided by the developed technologies. In the project, several applications will be assessed in key domains where Europe is strong (automotive, digital farming, digital industry, mobility and digital life). The aim is to reinforce and maintain strong leadership in these areas by bringing industry in contact with future memory technologies at a low TRL level (MRAM, OXRAM, FeFET). These cross-disciplinary efforts will lead to development of innovative hardware / software deep learning solutions, based on high TRL level RRAM/PCM and FeFET, to enable future products which combine extreme power efficiency with robust cognitive computing capabilities. This new kind of computing technology, combining ANN and SNN capabilities, will open new perspectives, for instance, environmental monitoring, and wearable electronics.

Serving the €17bn microcontroller industry: Every IoT device needs a microcontroller (MCU). Each MCU has logic transistors for processing data and eNVM (flash) for storing data, among other components. Most microcontrollers today are manufactured at 40nm nodes or above. Smaller technology nodes, like 28nm, promise lower power and lower cost, critical for IoT adoption. Logic transistor scale down fine to 28nm and below. But Flash does not scale down to smaller technology nodes – the effects of Moore’s law have hit a physical limit at 28nm for Flash memory; writing data will destroy them. Consequentially, the benefits of 28nm will be denied to MCU vendors, unless a new eNVM technology is found. FeFET, the ferroelectric field effect transistor by FMC, is based on the discovery that hafnium oxide, the gate material of modern logic transistors, can be modified such that it becomes ferro-electric. By applying a positive or negative voltage, the material can store a 0 or 1. The gate remembers the state even when power is turned off – the transistor becomes a non-volatile memory cell. As hafnium oxide retains ferro electricity to at least 2nm film thickness, we have clear path to eNVM cells even at the 7nm node – the most advanced technology node available today. In Phase 1, FMC wants to better understand the technical requirements in various sectors of the microcontroller market, develop a communication approach, and develop a detailed plan how to meet their application specific functional and performance requirements. Today, FMC is focused on advancing its technology from TRL5 to TRL6 to satisfy minimum viable product properties ahead of commencing prototype production. In Phase 2, FMC will use the base technology to create a range of custom-specific memory macros that satisfy the requirement of different market segments and customer types (TRL7-8). The company should reach more than 115M€ revenues by 2024 provided it can raise the required financing.

Information-and-communication technologies have been historically powered by silicon, with development and production taking place mostly, albeit not exclusively, in the United States and in Asia. The current and major worldwide drive for big data, machine learning, and quantum computing will push away from this all-silicon platform, and provide a unique opportunity and a clean slate for European industry to rapidly deploy novel technologies based on innovative materials and devices. Leadership will require fast exploration of materials’ properties (e.g. memory effects for memristive computing), linking properties to performance in unexplored architectures, and assessing their business potential. INTERSECT wants to leverage European leadership in materials’ modelling software and infrastructure, as embodied in track record of the team, to provide industry-ready integrated solutions that are fully compliant with a vision of semantic interoperability driven by standardized ontologies. The resulting IM2D framework - an interoperable material-to-device simulation platform - will integrate some of the most used open-source materials modelling codes (Quantum ESPRESSO and SIESTA) with models and modelling software for emerging devices (GinestraTM) via the SimPhony infrastructure for semantic interoperability and ontologies, powered by the AiiDA workflow engine, and its data-on-demand capabilities and apps interface. API-compliance with established standards will allow pipelines to and from public repositories, and embedding into the front-end of materials hubs, such as MarketPlace, while testing, validation, and standardization will take place together with the industrial partners. INTERSECT will drive the uptake of materials modelling software in industry, bridging the gap between academic innovation and industrial novel production, with a goal of accelerating by one order of magnitude the process of materials’ selection and device design and deployment.

The Ferro4EdgeAI project will provide an ultra-low power, scalable edge accelerator for artificial intelligence incorporating a memory augmented neural network, based on low cost, high density, multi-level, Back End of Line (BEoL) integrated ferroelectric (FE) technology. We expect to achieve a 2500x gain in energy-efficiency to break the POPS/W barrier with respect to the state-of-the-art CMOS accelerators and predictions for other emerging technology AI hardware. To do so, five ambitious specific objectives have been selected: - multi-level functionality in hafnia-based thin films by investigating the optimum trade-off in memory window, film thickness & stability of the ferroelectric state - low operating voltage for the non-volatile memory and robust multilevel operation of the FeFET-2 for high density logic operations and data storage. A low operating voltage is mandatory for power rating reduction, while robust multilevel operation is essential for analogue in-memory computing at the edge. - integration and characterization of multi-level, low voltage, FeFET-2 arrays - definition, design and demonstration of a low power FE AI accelerator suitable for scalable systems integration - Systems simulation of ultra-low power FE accelerator enhanced edge processing for targeted edge applications of voice and image recognition Ferro4EdgeAI is a multidisciplinary project engaging 12 partners from 6 countries covering the academic and industrial worlds (including 2 SMEs). An implementation plan is presented in the form of 6 work packages, 5 of which are technical in nature. Synergy in communication and dissemination by the several partners and stakeholders (including an external advisory board and collaboration with South Korea) will maximize the project progress and impact. Solutions to overcome the fundamental technological barriers as well as appropriate deliverables, tasks, milestones, and risks to complete the project objectives in due time are presented.

The semiconductor industry is a 600B$ industry. The share of semiconductor memories is 150B$. But there are no memory manufacturers in Europe. So, the EU is missing out on a huge share of the semiconductor market! FMC will change that – by creating fundamentally new memories based on our disruptive material innovation, we will become a major memory player from Europe, reducing the EU’s critical dependence on overseas suppliers, creating jobs, and making semiconductors greener. Our memory is as fast and low cost as DRAM, while providing non-volatile data storage. Our memory products will significantly reduce the energy consumption of data centers, enable instant-on capabilities, and double the battery life of smartphones, PCs, and IoT devices. In addition, non-volatile working memory immediately protects against data loss due to power outages. Our vision is to disrupt the memory industry and become the first EU memory company, set up production facilities in Europe to become the next EIC unicorn.