Why Magnetic Skyrmions Matter for Neuromorphic Computing as AI Energy Demand Grows
Neuromorphic computing, which aims to process information more like the human brain, is drawing new attention as AI’s energy needs climb. Magnetic skyrmions are one candidate building block for that approach. A 2025 study from Johannes Gutenberg University Mainz, now back in circulation, shows the kind of precise control researchers need before skyrmion-based devices become practical.
Key Takeaways
- A study published in Nature Nanotechnology in 2025 filmed a two-dimensional skyrmion lattice melting in real time while researchers controlled individual skyrmions.
- The research was led by Johannes Gutenberg University Mainz, with contributions from the Norwegian University of Science and Technology (NTNU), which recirculated the findings on October 26, 2026.
- The lattice melted in two distinct steps, unlike three-dimensional solids such as ice.
- Skyrmions are studied as possible building blocks for future data storage and brain-like computing.
- The International Atomic Energy Agency expects nuclear power use could more than triple by 2060, citing AI data centers among the drivers.
Neuromorphic Computing Aims to Work More Like the Brain
Conventional computers keep memory and processing in separate places. Data constantly moves back and forth between them, and that movement costs time and energy. Neuromorphic computing takes a different approach, modeled on how the brain’s neurons and synapses work. Memory and processing sit in the same place, and many operations happen in parallel.
The goal is computing that handles certain tasks, such as pattern recognition, with far less energy than today’s chips. Building that kind of hardware requires new physical components that can store and process information in different ways. Skyrmions are one of the options researchers are studying.
What Skyrmions Are and Why Researchers Study Them
Skyrmions are tiny magnetic vortices that form in thin magnetic materials, often described as microscopic magnetic whirlpools. They are stable, they can be moved with small amounts of energy, and they can be created and erased. Those properties make them interesting for spintronics, a field of electronics that uses the spin of electrons, not just their electric charge, to carry information.
Magnetic skyrmions can exist on their own, but when packed densely enough, they arrange themselves into an ordered lattice, much like atoms in a crystal. Understanding how that lattice forms and breaks down matters for any future device that would rely on skyrmions behaving predictably.
The Mainz Study Filmed a Skyrmion Lattice Melting in Real Time
The research behind this round of neuromorphic computing news was conducted mainly at the Institute of Physics at Johannes Gutenberg University Mainz, in Professor Mathias Kläui’s group, with lead author Raphael Gruber. Asle Sudbø, who heads NTNU’s Centre for Quantum Spintronics, contributed to the study. It was published in Nature Nanotechnology in 2025. NTNU’s October 2026 release has brought it back into wider circulation.
The team created skyrmions by carefully tuning temperature and magnetic fields, then watched them through a magneto-optical Kerr microscope, an instrument that images magnetic patterns using light. By adjusting magnetic fields, the researchers changed the size and mobility of individual skyrmions while filming. “For the first time, researchers have managed to film how a magnetic skyrmion lattice melts in real time,” Sudbø said in NTNU’s release.
The results showed that the two-dimensional lattice melts in two steps. First, the skyrmions lose their regular spacing while still holding a lattice pattern. Then the lattice’s orientation breaks down, and the structure dissolves completely. Shrinking the skyrmions made them more mobile, which pushed the lattice toward disorder. That is the kind of control over individual magnetic structures researchers will need if skyrmions are ever used to store or process data.
Energy Efficient Computing Is Becoming a Business Issue
The renewed interest in research like this reflects a practical problem: AI is expensive to power. In its annual report to the United Nations General Assembly, delivered October 23, the International Atomic Energy Agency said it expects nuclear power use could more than triple by 2060. The agency cited electrification, AI-powered data centers and technology companies’ demand for reliable, carbon-free power as drivers.
AI hardware spending is already showing up in economic data, including record capital goods imports tied to the AI infrastructure buildout. Every new data center needs power, and electricity costs flow through to cloud pricing and, eventually, to the businesses that rely on AI tools. Energy efficient computing has moved from a research goal to a cost concern for companies of every size.
Skyrmion-Based Computing Remains Early-Stage Research
The Mainz study is basic research. It explains how skyrmion lattices behave and how they can be controlled, but it does not describe a working device, a product or a timeline. Moving from controlled laboratory observations to reliable, manufacturable components is a long process, and many promising materials never make that leap.
The takeaway is directional rather than immediate. Research into fundamentally more efficient computing is active and well funded, and it is gaining relevance as AI’s energy demands grow. Skyrmions are one path among several, and studies like this one are the groundwork any eventual technology would depend on.
FAQs
What Is Neuromorphic Computing?
Neuromorphic computing is an approach to computer design modeled on the brain, where memory and processing happen in the same place and many operations run in parallel. The aim is to perform tasks such as pattern recognition using far less energy than conventional chips.
How Does Neuromorphic Computing Work?
Neuromorphic hardware uses components that mimic how neurons and synapses work, storing information where it is processed instead of moving data between separate memory and processing units. Researchers are testing several physical systems for this, including magnetic structures such as skyrmions.
What Is Spintronics?
Spintronics is a field of electronics that uses the spin of electrons, a magnetic property, in addition to their electric charge to store and process information. Magnetic skyrmions are studied within spintronics as possible building blocks for future memory and computing devices.
How Does AI Affect Energy Consumption?
AI runs in data centers that need large, constant supplies of electricity, which raises overall power demand. The International Atomic Energy Agency cites AI data centers among the drivers of its expectation that nuclear power use could more than triple by 2060.
