Research report: Building brain-inspired semiconductor chips for smarter, greener data centres

A systematic review by researchers at the University of Ottawa, Columbia University and Hanyang University examines emerging semiconductor materials and device architectures that could support faster, more energy-efficient brain-inspired computing. The review focuses on “post-CMOS” hardware neurons, an approach intended to address the substantial electricity demands of artificial-intelligence data centres while offering scalable artificial neurons for large-scale systems. Published in Advanced Functional Materials , the work evaluates advances across electronic materials, device physics and circuit integration rather than reporting a single experimental demonstration.

Lead author Dr. Kannan Udaya Mohanan, a postdoctoral fellow in Professor Chang-Hyun Kim’s research group at the University of Ottawa, said the materials and architectures represent a fundamentally different way of processing information. By operating more like a human brain than a traditional processor, the emerging technologies are expected to produce physical devices that are dramatically more energy-efficient, faster and adaptive. These are projections from the review, not demonstrated performance results for commercial data-centre systems.

The review addresses a limitation of conventional computing known as the “memory wall,” in which data transfer between memory and processors causes major delays and energy loss. According to the research team, the most promising post-CMOS materials and device architectures identified in the review can act as artificial neurons, with next-generation circuits designed to process information faster and with less energy than current mainstream chips. The authors mapped the broader hardware landscape to connect progress in materials science with pathways for developing scalable, brain-like systems.

Energy efficiency is only one of the potential advantages of these brain-inspired technologies. Post-CMOS refers to materials and architectures that move beyond traditional chips, and some of the emerging materials may be lower-cost, highly scalable and compatible with ultra-thin, lightweight and flexible form factors. The review therefore considers possible applications beyond greener AI data centres, including smarter, low-power wearable technology, personalised healthcare and human-machine interfaces.

These proposed uses extend the hardware roadmap into areas in which compactness, adaptability and reduced power consumption could be valuable. Taken together, the review presents hardware neurons as a possible response to the energy demands associated with AI data centres and the bottlenecks created by conventional processor-memory communication. Its central conclusion is that emerging post-CMOS materials, artificial-neuron devices and system integration offer a credible research direction for computing that is more energy-efficient, fast and adaptive.

Because the publication is a systematic review and forward-looking roadmap, its role is to identify promising technologies and integration pathways rather than to establish finished products or measured data-centre savings.

Subject of Research: A systematic review of post-CMOS hardware neurons for energy-efficient, brain-inspired computing systems

Article Title: Building brain-inspired semiconductor chips for smarter, greener data centres

Article References: Building brain-inspired semiconductor chips for smarter, greener data centres. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: neuromorphic computing, hardware neurons, post-CMOS devices, artificial intelligence, data centres, energy efficiency, semiconductors, artificial neurons, wearable electronics, human-machine interfaces, Building, brain-inspired