AI is evolving at an incredible pace, but its growing energy demands pose a major challenge. Enter spintronic devices—new technology that mimics the brain’s efficiency by integrating memory and processing.

Scientists in Japan have now developed a groundbreaking spintronic device that allows for electrical control of magnetic states, drastically reducing power consumption. This breakthrough could revolutionize AI hardware by making chips far more energy-efficient, mirroring the way neural networks function.

Spintronic Devices: A Game-Changer for AI Hardware

AI is rapidly transforming industries, but as these technologies evolve, so does their demand for power. To sustain further advancements, AI chips must become more energy efficient.

This is where spintronic devices come in. By integrating memory and computing functions—similar to how the human brain operates—they offer a promising foundation for low-power AI chips.

Now, researchers from Tohoku University, the National Institute for Materials Science, and the Japan Atomic Energy Agency have developed a groundbreaking spintronic device. This new technology enables the electrical mutual control of non-collinear antiferromagnets and ferromagnets, allowing for efficient switching of magnetic states. In practical terms, it can store and process information using significantly less energy, much like a brain-inspired AI chip.

This breakthrough could pave the way for a new generation of AI hardware that is both highly efficient and energy-saving. The findings were published in Nature Communications on February 5, 2025.

New AI Function Electrically Programmable Spintronic Device
A new spintronic breakthrough could cut AI’s power needs dramatically. Scientists have developed a way to switch magnetic states electrically, mimicking neural networks and enabling ultra-efficient AI chips. Credit: Tohoku University

Revolutionizing AI with Multi-State Magnetic Control

“While spintronic research has made significant strides in controlling magnetic order electrically, most existing spintronic devices separate the role of the magnetic material to be controlled and the material providing the driving force,” says Tohoku University’s Shunsuke Fukami, who supervised the research.

These devices have a fixed operation scheme once fabricated, typically switching information from “0” to “1” in a binary fashion. However, the new research team’s breakthrough offers a major innovation in electrically programmable switching of multiple magnetic states.

Electrical Mutual Switching Device
Schematic illustration of (a) a conventional magnetic memory device and (b) the device for electrical mutual switching developed in this work. Credit: ©Shunsuke Fukami

Harnessing the Power of the Magnetic Spin Hall Effect

Fukami and his colleagues employed the non-collinear antiferromagnet Mn3Sn as the core magnetic material. By applying an electrical current, Mn3Sn generates a spin current that drives the switching of a neighboring ferromagnet, CoFeB, through a process known as the magnetic spin Hall effect. Not only does the ferromagnet respond to the spin-polarized current, but it also influences the magnetic state of Mn3Sn, enabling the electrical mutual switching between the two materials.

In their proof-of-concept experiment, the team demonstrated that information written to the ferromagnet can be electrically controlled via the magnetic state of Mn3Sn. By adjusting the set current, they were able to switch the magnetization of CoFeB in different traces representing multiple states. This analog switching mechanism, where the polarity of the current can change the sign of the information written, is a key operation in neural networks, mimicking the way synaptic weights (analog values) function in AI processing.

Neuromorphic Computing Enabled by Electrical Mutual Switching
Proof-of-concept functionality for neuromorphic computing enabled by the phenomenon of electrical mutual switching. Credit: ©Shunsuke Fukami

Paving the Way for Energy-Efficient AI Chips

“This discovery represents an important step toward the development of more energy-efficient AI chips. By realizing the electrical mutual switching between a non-collinear antiferromagnet and a ferromagnet, we have opened new possibilities for current-programmable neural networks,” said Fukami. “We are now focusing on further reducing operating currents and increasing readout signals, which will be crucial for practical applications in AI chips.”

The team’s research opens new pathways for improving the energy efficiency of AI chips and minimizing their environmental impacts.

Reference: “Electrical mutual switching in a noncollinear-antiferromagnetic–ferromagnetic heterostructure” by Ju-Young Yoon, Yutaro Takeuchi, Ryota Takechi, Jiahao Han, Tomohiro Uchimura, Yuta Yamane, Shun Kanai, Jun’ichi Ieda, Hideo Ohno and Shunsuke Fukami, 5 February 2025, Nature Communications.
DOI: 10.1038/s41467-025-56157-6

Read The Article

News