Puyang Huang

About Puyang’s research on Energy-efficient MRAM for High-performance AI Computing:

The explosive growth of modern AI models has brought unprecedented demands for computing performance and energy efficiency. Traditional CMOS-based hardware struggles to balance speed, power, and scalability, motivating the exploration of emerging memory technologies for next-generation AI systems.

My research focuses on magnetic random-access memory (MRAM), particularly spin–orbit torque (SOT) MRAM and voltage-controlled magnetic anisotropy (VCMA) MRAM, which are promising candidates for energy-efficient and high-speed computing. By exploiting their non-volatility, high endurance, and sub-nanosecond switching capability, I investigate how MRAM devices can be integrated into in-memory computing, neuromorphic architectures, and edge AI platforms. Ultimately, my work aims to bridge device-level physics with system-level design, enabling intelligent computing that is both powerful and sustainable.