A major breakthrough in semiconductor technology brings us closer to next-generation artificial intelligence. Researchers have developed atomically tunable “memristors”—innovative memory resistors that mimic the human brain’s neural networks. This advancement could pave the way for neuromorphic computing, a cutting-edge method of processing data that mirrors the brain’s ability to learn, adapt, and operate efficiently.
A new era of neuromorphic computing
This project, funded by a US$1.8 million grant from the National Science Foundation’s Future of Semiconductors program (FuSe2), aims to revolutionise computing. By creating ultrathin memory devices with atomic-scale control, scientists are opening the door to AI systems capable of higher speed, efficiency, and adaptability. These devices, known as memristors, can act as artificial synapses and neurons, making them integral to brain-inspired computing.
The University of Kansas (KU) and the University of Houston are leading this effort, with a team led by Judy Wu, a distinguished professor of physics and astronomy at KU. The team has succeeded in developing memory devices as thin as 0.1 nanometres—approximately 10 times thinner than typical nanometre-scale components. This level of precision is vital for creating highly efficient and scalable semiconductors.
Memristors are particularly well-suited for neuromorphic circuits as they can store and process data simultaneously. This ability to handle parallel data streams mirrors the functioning of the biological brain. It could overcome the limitations of traditional computing systems, which struggle with the energy demands of modern AI applications.
Tackling challenges in computing
One of the central challenges of neuromorphic computing lies in achieving the precision and scalability needed for brain-like systems. The research team employs a co-design approach, integrating material design, fabrication, and testing to ensure their devices meet these requirements.
The project also seeks to address the growing demand for skilled professionals in the semiconductor industry. Through an educational outreach initiative, researchers hope to inspire and train the next generation of experts in this field. This project component is a collaboration between KU and the University of Houston, ensuring a comprehensive approach to workforce development.
Building a brain-inspired future
“The overarching goal of our work is to develop atomically ‘tunable’ memristors that can act as neurons and synapses on a neuromorphic circuit,” said Wu. “We aim to mimic how our brain thinks, computes, makes decisions, and recognises patterns—all with high speed and energy efficiency.”
This breakthrough promises significant advancements in AI, enabling systems that learn and adapt like the human brain. Such capabilities could transform industries reliant on machine learning, from healthcare to robotics.
As research progresses, these ultrathin, brain-like circuits could mark a pivotal step towards smarter, more sustainable computing solutions, changing how we interact with technology and the world around us.
