AI ‘Memory Bottleneck’ Could be Solved by New Technology

An optical image of the device structure with 4 micrometer pillar diameter. Image Credit: Northwestern University/University of Messina, Italy

A team of engineers from Northwestern University in the US and the University of Messina in Italy has established a system that could overcome the ‘memory bottleneck’ that currently limits the capabilities of existing computing hardware, which cannot keep up with the rapid advancements in data-centric computing.


Using antiferromagnetic (AFM) materials, the international team developed the world’s smallest device of this kind with the capabilities of writing data using a record-low electrical current. This magnetic memory device is considered to be a viable solution to the ever-increasing need for increased power, storage, and speed, as data-centric computing continues to evolve at rapid rates.


Hardware Limits Potential of Data-Centric Computing

Recent years have seen the evolution of artificial intelligence (AI) out of developments being made in the field of big data. This establishment of AI is infiltrating all industries and is fundamentally changing how we use technology. It’s drastically advanced how we create networks and store data, however, as this side of the technology continues to evolve, it is facing limitations posed by the hardware that is being used to support it. Currently, hardware that was built for the previous generation of computing is still being relied upon to sustain the boom in data-centric computing, but its storage, power, and speed confines are preventing data-centric computing from reaching its full potential.


The technology that has been developed by the US/Italian team proposes a solution to this key industry challenge. In a paper published this month in the journal Nature Electronics, the team describes how they innovated their device with the help of antiferromagnetic (AFM) materials.


Storing Data In Magnetic Fields

For AI to work to its full potential it requires a system that integrates all the capabilities of the latest technology in memory, such as the speed of static random access memory (SRAM), the storage capacity of dynamic random access memory (DRAM) or Flash, along with incredibly low power dissipation.


The engineering team recognized that there was no existing technology that integrated all of these capabilities into one single system. The impact of this was that AI applications were limited by what was termed as a ‘memory bottleneck’.


The team aimed to address this limitation, and AFM materials were looked to, to provide a solution. The electrons within AFM materials have particular properties that make them useful for designed next-generation technologies. Due to “spin”, a quantum mechanical property, the atoms in AFM materials act like magnets. This means that while the bulk material does not have macroscopic magnetization, as the spins of the atoms are aligned in antiparallel, the atoms demonstrate magnetically ordered spins. It is this feature that has made them desirable for creating next-generation electronics.


An electrical current is generally required for the storage of data in classical memory devices. AFM materials, however, replace the reliance on an electrical current, using their magnetically ordered spins to retain the stored data instead. This eliminates the need for a continuous supply of electrical current. Also, AFM-based devices do not interact with magnetic fields because of their densely packed electrons, meaning that the devices can be designed to be both incredibly small and secure, with external, magnetic fields having no impact on the stored data.


Future Goals

While previous research exploring into the use of AFM materials in this capacity has failed to control the magnetic order within the materials, the current findings demonstrate a system that has overcome this problem and has developed AFM materials for use in creating a successful data storage device that is fast and secure, with low power demands.


What’s important in what was achieved by the US/Italian team is that the device they established is easily compatible with semiconductor manufacturing practices that are already established within the industry. Therefore, the technology could be adopted by manufacturing companies without the need for major investments in new equipment.


The next steps will be working on making the devices smaller, and discovering more energy-efficient techniques of writing data into the AFM materials. Once these goals have been reached we can expect these devices to become widely adopted, likely having a significant impact on the sectors of security, healthcare, and transportation, to name a few.

Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.

Sarah Moore

Written by

Sarah Moore

After studying Psychology and then Neuroscience, Sarah quickly found her enjoyment for researching and writing research papers; turning to a passion to connect ideas with people through writing.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Moore, Sarah. (2020, February 11). AI ‘Memory Bottleneck’ Could be Solved by New Technology. AZoRobotics. Retrieved on February 29, 2024 from

  • MLA

    Moore, Sarah. "AI ‘Memory Bottleneck’ Could be Solved by New Technology". AZoRobotics. 29 February 2024. <>.

  • Chicago

    Moore, Sarah. "AI ‘Memory Bottleneck’ Could be Solved by New Technology". AZoRobotics. (accessed February 29, 2024).

  • Harvard

    Moore, Sarah. 2020. AI ‘Memory Bottleneck’ Could be Solved by New Technology. AZoRobotics, viewed 29 February 2024,

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type
Azthena logo powered by Azthena AI

Your AI Assistant finding answers from trusted AZoM content

Your AI Powered Scientific Assistant

Hi, I'm Azthena, you can trust me to find commercial scientific answers from

A few things you need to know before we start. Please read and accept to continue.

  • Use of “Azthena” is subject to the terms and conditions of use as set out by OpenAI.
  • Content provided on any AZoNetwork sites are subject to the site Terms & Conditions and Privacy Policy.
  • Large Language Models can make mistakes. Consider checking important information.

Great. Ask your question.

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.