February 25, 2026 by Sam Jarman, Phys.org
Collected at: https://phys.org/news/2026-02-2d-memristors-ai-energy-problem.html
New generations of memristors could reliably store information directly within the molecular structures of graphene-like materials. In a new review published in Nanoenergy Advances, Gennady Panin of the Russian Academy of Sciences shows how these atomically thin materials are ideally suited for electrical circuits that mimic the function of our own brains—and could help address the vast power requirements of emerging AI technologies.
Remembering past currents
A memristor is a cutting-edge electrical component whose resistance depends on the amount of current that previously passed through it. Because it “remembers” this history even after charge is no longer flowing, it can store data when the power is switched off. In this way, memristors operate in a way remarkably similar to the neurons in our brains and the synapses connecting them.
With their fast response times, combined with simple, two-electrode structures that allow them to be packed into dense arrays, memristors are increasingly forming the building blocks of modern circuits—especially those designed for AI.
Graphene-based materials
In his review, Panin explores how these capabilities could be pushed even further by building memristor circuits from 2D materials just a few atoms thick.
Since the discovery of graphene, numerous studies have examined how its already versatile electrical properties can be enhanced by modifying the molecular structure of its honeycomb lattice of carbon atoms.
In particular, these materials can be engineered to exhibit nonlinear behavior—meaning the current flowing through them doesn’t increase proportionally with applied voltage. This nonlinearity is essential for stable memory storage and switching between distinct resistance states.
Controllable memristive states
Panin considers a versatile range of 2D materials: including graphene oxide, diamane (a 2D, diamond-like phase of carbon), and layered chalcogenides—which don’t contain carbon but share a similar structure to graphene.
Across these materials, electrical current can trigger partial rearrangements of their atomic lattices—shifting from flat, highly conductive networks to more distorted, less conductive configurations, increasing their resistance. In graphene-based systems, these effects can also be tuned through controllable redox reactions: adding oxygen-containing groups makes the material less conductive, while removing them restores higher conductivity.
Finally, the review explores how memristive properties can also emerge through phase transitions triggered by light across a broad range of wavelengths. Such optically driven switching enables devices that both sense and store information—comparable to how living systems gather and retain information from the light they perceive.
Possibilities for AI
By highlighting these robust switching mechanisms, Panin ultimately argues that 2D graphene-like materials could help address the rapidly growing energy demands of AI data centers—currently one of the most pressing concerns surrounding the technology’s rapid expansion.
By integrating memory storage directly into the molecular structure of circuit elements, such devices could perform similarly fast and powerful calculations to existing architectures while consuming only a fraction of the energy. In turn, this approach could pave the way for more sustainable AI applications: from self-driving cars to the discovery of new medicines tailored to individual patients.
More information
Gennady N. Panin, Graphene-Based Memristive and Photomemristive Nanosensors for Energy-Efficient Information Processing, Nanoenergy Advances (2026). DOI: 10.3390/nanoenergyadv6010006
Leave a Reply