Flexoelectricity: How 100 Trillion Cells Could Inspire AI

Every day, your body carries around 100 trillion tiny batteries. That is enough to spark some wild ideas among computer engineers and AI researchers alike.

A particularly electrifying finding was published on 12 December 2025 in the journal PNAS Nexus. Researchers from the University of Houston and Rutgers University in the United States examined how the cells that make up our bodies operate, with a particular focus on the lipid membranes surrounding them. These membranes act as a protective barrier: they help cells maintain their integrity when exposed to external threats, such as pathogens, filter the molecular information received by the body, and regulate metabolic exchanges.

Yet that is not their only role. These same membranes may be able to convert their natural ripples into electricity, much like a generator. This previously unknown energy source could eventually help create bio-inspired electronic devices.

Flexoelectricity: converting movement into voltage

Cells produce this energy through flexoelectricity, a physical phenomenon in which electricity arises from a change in shape. Picture a cell membrane as a sheet of A4 paper: sharply twisting and bending it creates a separation of electrical charges between the inner and outer sides of the bend.

At cellular scale, this mechanism is continuous because membranes are constantly rippling under the effects of heat and the activity of proteins passing through them. As they deform thousands of times every second, they effectively act as electromechanical nanogenerators. The researchers behind the study explain: “We show here that these active fluctuations, when coupled to the universal property of flexoelectricity, can generate transmembrane voltages and even direct ion transport.”

Through their membranes, cells harvest the energy required to power their biological activity. The team’s calculations suggest that these tiny ripples could produce a potential difference of up to 90 mV (millivolts). Although that figure may seem negligible compared with a mains socket - around 2,500 times lower - it is enormous on a cellular scale. It represents enough charge to trigger a neuron's nerve impulse or control the contraction of a muscle fibre.

A fresh source of inspiration for AI and technology

This discovery is naturally of great interest to every profession linked by technology. The energy harvesting carried out by our cells could one day contribute to a new generation of smart materials and biomimetic innovations. One possibility would be to develop more energy-efficient artificial intelligence networks - currently a major weakness - by replacing transistors in electrical circuits with bio-inspired nanogenerators that would power themselves using vibrations from their surroundings.

Such a process might perhaps one day be applied to processors, which also require substantial amounts of energy to process information. Cells are the opposite: they handle complex tasks partly by recycling fluctuations in their membranes. The study’s authors believe that: “Investigating electromechanical dynamics in neural networks could bridge molecular flexoelectricity and complex information processing.”

It is important to remain realistic, however: using flexoelectricity to power even a micro-electrical system is not yet within our reach. For now, this is only a theoretical model, and it will still need to be tested in vivo to determine whether it is truly valid. That does not make the study any less significant, as it is the first to show that our cells rely as much on mechanics as on organic chemistry to generate their energy flow. Smartphones or watches powered by “biological batteries” may exist one day, but a great deal of time will pass before that becomes reality.