HKU Engineers Build 10mK Neuromorphic Chip for Quantum Computing and Deep-Space Missions
Updated
Updated · hku.hk · Jun 7
HKU Engineers Build 10mK Neuromorphic Chip for Quantum Computing and Deep-Space Missions
1 articles · Updated · hku.hk · Jun 7
Summary
HKU researchers developed a programmable neuromorphic hardware platform that operates at 10 millikelvin, letting a single silicon-carbide transistor mimic neuron-like spiking near absolute zero.
The advance targets a key quantum-computing bottleneck: today’s silicon controllers run too hot and power-hungry to sit near qubits, creating wiring limits that constrain scaling and performance.
Below 2 kelvin, the team found industry-standard SiC MOSFETs show stable S-shaped negative differential resistance driven by electron-donor impact ionization, enabling circuits they said are thousands of times more energy-efficient than conventional electronics.
Because SiC is already manufactured at scale for electric vehicles and power grids, the chips could be produced in existing foundries on 300-mm wafers and cascaded into larger cryogenic neural networks.
Published in Nature Communications, the work could support quantum error correction and real-time control, while also offering rugged electronics for the extreme cold of the Moon and deep space.
A material known to fail in extreme cold is now a quantum breakthrough. What is the secret, and is the technology truly reliable?
Will a breakthrough in electric car chip material finally unlock the power of quantum computers, or do critical manufacturing hurdles remain?
World's First 10 Millikelvin Programmable Cryogenic Neuromorphic Hardware: SiC MOSFET Breakthrough Paves Way for Quantum Computing and Deep-Space AI
Overview
The University of Hong Kong has achieved a world-first by unveiling a programmable neuromorphic hardware platform that operates at an ultracold 10 millikelvin. This breakthrough uses Silicon Carbide (SiC) MOSFETs, which allow the hardware to mimic the energy-efficient 'spiking' behavior of biological neurons. As a result, these advanced chips can overcome the high energy consumption and computational limits of traditional computers. Since SiC technology is already widely used in industries like electric vehicles and power grids, existing manufacturing processes can be leveraged to produce these cryogenic chips at scale, making this innovation both practical and impactful.