Imagine a computer chip that remembers information and processes it in the same spot, like a human brain, instead of moving data back and forth. This project uses special materials that act like tiny switches to make AI run on almost no power. It's like replacing a gas-guzzling engine with a high-efficiency electric motor for the 'brains' of smart devices.
Current AI hardware consumes too much energy and wastes power moving data between memory and processors. This prevents the deployment of truly self-sustaining, miniature AI sensors.
The project is developing scaled ferroelectric synaptic devices (<20nm) and ultra-dense crossbar arrays for non-volatile multi-bit digital functionality.
If you are a wearable device manufacturer dealing with short battery life in AI health trackers — this project developed ferroelectric synaptic devices that operate in the <20nm regime to drastically reduce power consumption.
If you are a smart sensor producer dealing with high energy costs for edge-computing — this project developed ultra-dense crossbar arrays that allow for highly parallel calculations directly on the chip.
If you are an autonomous vehicle component supplier dealing with the heat and energy load of AI processing — this project developed CMOS compatible materials that minimize data movement between storage and computing units.
Based on available project data, there is no specific pricing or cost-per-unit information provided.
The project focuses on scaling devices to the <20nm regime and uses CMOS compatible materials, which are designed for integration into existing semiconductor manufacturing processes.
Based on available project data, specific licensing terms are not listed, but the project leverages European discoveries in HfO2 and AlScN materials.
The technology is designed for hybrid integration and packaging together with conventional CMOS designs.
The project is active from 2023-11-01 to 2027-04-30.
The consortium is well-balanced for hardware development, consisting of 12 partners across 5 countries. It features a 33% industry ratio (4 industrial partners), complemented by 4 universities and 4 research organizations, ensuring a pipeline from fundamental material science to industrial application.
Contact NaMLab GGMBH in Germany
Contact us to track the TRL progression of ferroelectric AI hardware.
