StorAIge - New storage technology for edge AI applications

StorAIge - New storage technology for edge AI applications

Project period: 07/2021-06/2024

Artificial intelligence (AI) is being used in more and more applications today and is already considered a key technology for key technology for electronic components and systems. Today's AI technologies are inefficient and expensive and often still suffer from low acceptance among the general public. In the StorAIge project, Fraunhofer IPMS has joined forces with European partners to develop a platform for silicon-based AI chips that is high-performance, energy-efficient, and secure to enable competitive edge AI applications applications in the automotive, industrial, security, and consumer sectors.

Most AI systems today are used for data analysis and data-driven decision-making. Edge AI approaches, where intelligent data analysis is performed directly on the computer chip, offer major advantages in this regard: Decisions can be made faster to be (lower bandwidth usage), large amounts of data do not have to be shuttled back and forth across a network (lower data storage). This reduces overall system power consumption. At the same time, since data is processed locally, security increases. 

The main challenge of the project is, on the one hand, to manage the complexity of "More than Moore" technologies smaller than 28 nm and bring them to a high level of maturity and, on the other hand, to develop the design of complex systems-on-chip for smarter, safer, more flexible, low-power and lower-cost applications. The project targets chips with very efficient memories and high computing power to reach 10 tops per watt. To make this possible, Fraunhofer IPMS is relying on ferroelectric field-effect transistors (FeFETs). These are implemented directly on the semiconductor chip. This allows Fraunhofer IPMS to extend its experience in integrating, characterizing and optimizing ferroelectric memory technology on a scaled CMOS platform (latency) and without relying on network connections.


Ali, F., Ali, T., Lehninger, D., Sünbül, A., Viegas, A., Sachdeva, R., Abbas, A., Czernohorsky, M., Seidel, K.

Fluorite-Structured Ferroelectric and Antiferroelectric Materials: A Gateway of Miniaturized Electronic Devices

Adv. Funct. Mater. 2022, 2201737.