Over the last decade, with the introduction of increasingly complex artificial intelligence (AI) technologies, the demand for computing power has risen exponentially. New, energy-efficient hardware designs could help meet this demand while reducing computing’s energy use, supporting faster processing, and allowing AI training to take place within the device itself.

“In my opinion, we have already transitioned from the internet era to the AI era,” says EE Professor Shan Wang. “We want to enable AI on edge – training locally on your home computer, phone, or smartwatch – for things like heart attack detection or speech recognition. To do that, you need a very fast, non-volatile memory.”

Shan and his colleagues recently found a material that could bring a new type of memory closer to commercialization. Their paper, “Observation of anti-damping spin–orbit torques generated by in-plane and out-of-plane spin polarizations in MnPd3,” published in Nature Materials, the researchers demonstrated that a thin layer of a metallic compound called manganese palladium three had the necessary properties to facilitate a form of working memory that stores data in electron spin directions. This method of memory storage, known as spin orbit torque magnetoresistive random access memory or SOT-MRAM, has the potential to store data more quickly and efficiently than current methods, which store data using electric charge and require a continuous power input to maintain that data.

Shan reports, “We’ve provided a basic building block for future energy-efficient storage elements. It’s very foundational, but it’s a breakthrough.”

Excerpted from Stanford News, "A new material could enable more efficient magnet-based computer memory.”