The microscopic magnetic flips behind digital memory could soon use thousands of times less energy, offering a new way to shrink AI’s rapidly growing power footprint.
Artificial intelligence is creating and processing data on an enormous scale. Searches, recommendations, generated images, scientific simulations, and large language models all depend on information that must be repeatedly stored, transferred, retrieved, and rewritten. Each operation consumes energy, and those costs multiply across the billions of devices and sprawling data centers that support the digital world.
Researchers at the University of Edinburgh have now developed a mathematical framework designed to slash the energy required to write information in future magnetic memory. Rather than creating a new memory material, the approach changes how the magnetic state representing a digital bit is flipped.
Magnetic memory stores information by controlling the orientation of tiny magnetic regions. Switching one of these regions between two stable states can represent changing a bit from a “0” to a “1,” or the reverse. The challenge is to complete that switch quickly, reliably, and with as little wasted energy as possible.
Conventional magnetic field pulses are often too inefficient for modern memory architectures. They can consume substantial energy, struggle to target extremely small regions, and become difficult to scale as devices shrink. These limitations have pushed the industry toward technologies that use electrical currents, including spin-transfer torque magnetic random access memory (STT-MRAM) and spin-orbit torque magnetic random access memory (SOT-MRAM).
A Smarter Way to Flip Bits
The Edinburgh team revisited magnetic field switching using optimal control theory, a mathematical method for finding the most efficient route to a specific result. Instead of applying a simple pulse with a fixed shape, the framework calculates how the magnetic field should change from moment to moment to guide the magnetization into its new state with minimal energy.
This is similar to finding the most economical path through a complicated landscape rather than simply applying more force. The carefully shaped pulses take advantage of the magnet’s natural motion, steering it toward the desired state instead of fighting against its dynamics.
Ultrafast Switching With Far Less Energy
The researchers tested the method in computer simulations of three ultrathin van der Waals magnets: Fe₃GaTe₂, Fe₃GeTe₂, and CrSBr. These materials consist of weakly bonded atomic layers and are being investigated for compact, high-speed electronics because their magnetic properties can be controlled at extremely small scales.
In the simulations, the optimized pulses reversed the magnetization in about 1 to 10 picoseconds. One picosecond is one trillionth of a second. The fields were also more than 10 times weaker than those used in standard switching methods. Under the conditions studied, switching required as little as 0.94 nanojoules, compared with as much as 91.2 nanojoules for conventional field pulses.
By adjusting properties such as magnetic damping and anisotropy, which determine how spins move and which direction they prefer to point, the calculations suggest that switching energy could eventually fall into the femtojoule range. A femtojoule is one quadrillionth of a joule. At those levels, the approach could compete with or outperform established STT and SOT switching techniques.
Approaching Computing’s Energy Limit
The broader modeling indicates that optimized switching could reduce energy use by several orders of magnitude compared with leading memory technologies, including DRAM, STT-MRAM, and emerging SOT-MRAM devices. The predicted values also move magnetic memory closer to the Landauer limit (the fundamental thermodynamic limit), which defines the smallest amount of energy that must be dissipated when information is irreversibly erased.
The study, published in Advanced Materials, describes possible device designs and methods for delivering the carefully controlled fields, providing a starting point for experimental testing.
Dr. Elton Santos of the University of Edinburgh’s Institute for Condensed Matter Physics and Complex Systems, who led the research, said: “Every digital operation has an energy cost, and that cost becomes increasingly important as AI and data-intensive technologies continue to expand. Our work shows that, by carefully designing how a magnetic field changes in time, magnetization can be switched far more efficiently than with conventional approaches.”
Beyond Magnetic Field Pulses
The method may also extend beyond magnetic fields. The same mathematical strategy could be used to optimize the electrical currents and ultrafast laser pulses being explored for future memory and spintronics devices. That versatility could allow engineers to combine fields, currents, and light rather than relying on a single switching mechanism.
Santos continued: “Although we first developed the theory using magnetic field pulses, the mathematics is far more versatile than that. The same framework can be adapted to electrical currents and even ultrafast laser pulses, which are among the most cutting-edge technologies for future data storage. That means the ideas developed here could have applications far beyond the systems we studied. It seems that we may have just found the next best thing.”
Reference: “Optimal Control Drives Ultrafast and Energy-Efficient Magnetization Switching in Van der Waals Magnets” by Mohammad H Badarneh, PeiYu Cai and Elton J G Santos, 14 July 2026, Advanced Materials.
DOI: 10.1002/adma.202523059
Never miss a breakthrough: Join the SciTechDaily newsletter.
Follow us on Google and Google News.
3 Comments
Does this mean they’ll stop buying RAM and I can afford a new computer?
“ have now developed a mathematical framework designed to slash the energy required …”
So nothing has actually happened and almost certainly will not.
Across aerospace, thermal dynamics, power electronics, photonics, and computing architecture, these six discoveries demonstrate a single universal principle: peak energy efficiency is achieved by replacing rigid boundary resistance with organized, low-drag micro-geometry.
Mainstream engineering consistently hits operational limits when forcing energy through static, non-aligned materials:
Aircraft waste massive fuel fighting turbulent boundary-layer drag.
Power electronics breakdown when forced into mismatched crystal structures.
Computing platforms burn thousands of times more energy simply shuffling data across physical transport gaps between memory and processing units.
Each breakthrough overcomes a historic efficiency barrier by aligning system mechanics with natural, low-resistance pathways:
Aerospace & Thermal Flow: Micro-vibrations on aircraft wings and nanoscale polymer nucleation sites on heat exchangers replace chaotic, static turbulence with synchronized, low-impedance transport.
Semiconductor & Memory Architecture: In-memory computing arrays and SiC bottom-gate JFETs eliminate transport distance and lattice mismatch, operating directly within native physical structures.
Photonics & Material Science: Frequency combs and relaxor ferroelectrics replace brute-force resistance with continuous, parallel energy distribution.
Whether managing airflow over a wing, thermal transport across a copper surface, or data flow through a silicon chip, the path to next-generation performance relies on organizing micro-scale geometry to eliminate systemic drag. The historic “von Neumann bottleneck” isn’t a computing limitation—it is a spatial routing penalty from moving data back and forth across physical gaps. Performing calculations directly inside the memory lattice where data resides aligns state storage and state execution in the exact same spatial location. Eliminating transport distance eliminates transport impedance.