Researchers Synchronize 105,000 Nano-Oscillators in 45 Nanoseconds for Unconventional Computing
Updated
Updated · Nature.com · Jul 10
Researchers Synchronize 105,000 Nano-Oscillators in 45 Nanoseconds for Unconventional Computing
1 articles · Updated · Nature.com · Jul 10
Summary
A lattice of up to 105,000 spin Hall nano-oscillators reached full mutual synchronization, with phase ordering taking about 45 ns in the largest array and 10 ns in 100-oscillator arrays.
Microwave measurements showed coherence improving with scale: output power rose linearly with oscillator count, linewidth fell as N^-1, and one array achieved a quality factor above 10^6 at 26.2 GHz.
Time-resolved Brillouin light scattering found synchronization time increased only weakly—roughly logarithmically—with array size, matching Kuramoto-type collective ordering in a two-dimensional oscillator lattice.
The arrays used 10-20 nm constrictions, tight 24-40 nm spacing, and a W-Ta/CoFeB stack designed to cut power loss and improve heat removal, enabling stable room-temperature operation.
The result extends synchronized SHNO networks from prior demonstrations of 64 oscillators to more than three orders of magnitude larger, supporting prospects for embedded-Ising and reservoir-computing hardware.
As nano-oscillators solve problems in nanoseconds, will they outpace quantum computers for real-world optimization tasks?
This breakthrough mimics brain dynamics on a small scale; what are the limits to building a truly brain-like computer?
Can this nanosecond chip breakthrough shatter the 'Memory Wall' that is throttling the progress of modern AI?
Record-Scale Synchronization of 105,000 SHNOs in 45 ns: Implications for Neuromorphic and Analog Computing
Overview
In July 2026, a groundbreaking study published in Nature Nanotechnology reported the successful mutual synchronization of 105,000 spin Hall nano-oscillators (SHNOs) within just 45 nanoseconds. This achievement marks the largest and fastest synchronized network of its kind, setting new records for stability and precision with a quality factor of 18,000. The breakthrough paves the way for new forms of energy-efficient computing hardware, demonstrating exceptional control and efficiency in large-scale nano-oscillator networks and opening exciting possibilities for unconventional computing architectures.