Beyond Memory: MRAM as the Computing Engine for AI in the Co-Design Era

4:20 pm - 4:40 pm

As AI systems continue to scale, data movement between processors and memory has emerged as a fundamental bottleneck in energy consumption, latency, and system efficiency. This challenge is driving a paradigm shift from conventional von Neumann architectures toward memory-centric computing frameworks, where memory devices actively participate in computation.

Among emerging memory technologies, Magnetoresistive random access memory (MRAM) offers a unique combination of non-volatility, high endurance, ultrafast operation, CMOS compatibility, and intrinsic magnetic dynamics, making it a strong candidate for next-generation AI hardware. Beyond data storage, MRAM can support in-memory computing, neuromorphic functions, and temporal information processing through physics-enabled device behaviors.

This talk will present the evolution of Magnetoresistive Random Access Memory (MRAM) from conventional non-volatile memory toward neuromorphic and compute-in-memory architectures for AI hardware. Recent advances in spintronic devices, including analog switching, synaptic plasticity, leaky-integrate-fire neurons, and SOT-MRAM-based computing primitives, will be highlighted to demonstrate how device physics can directly enable computational functionalities. The presentation will further discuss how cross-layer co-design spanning materials, devices, circuits, and system architectures is transforming MRAM from conventional memory into an active computing platform for energy-efficient AI systems.

Key Technologies Covered

Spintronic Memory Technologies (MRAM)
Spin-Orbit Torque MRAM (SOT-MRAM)
Embedded Non-Volatile Memory for AI Hardware
Compute-in-Memory (CIM) Architectures
Neuromorphic Computing Hardware
Artificial Synapses and Leaky-Integrate-Fire Neurons

Featured Speakers

Prof. Chih-Huang Lai

Associate Dean, National Tsing Hua University

Prof. Lai received his Ph.D. in Materials Science and Engineering from Stanford University in 1997 and joined National Tsing Hua University (NTHU) the following year. His pioneering research has advanced the development of novel materials for memory technologies, renewable energy, and sustainable materials systems, with landmark contributions in magnetic random-access memory (MRAM), thin-film solar cells, and materials recycling.

He has published more than 240 research papers and holds over 50 awarded patents. His work has not only deepened fundamental scientific understanding but has also driven significant industrial applications. In recognition of his outstanding achievements, Prof. Lai has received numerous prestigious honors, including the Asian Union of Magnetics Societies (AUMS) Award, election as an Academician of the Asia Pacific Academy of Materials (APAM), IEEE Fellow, MRS-T Fellow, and the Micron Chair Professorship.

Currently, Prof. Lai serves as the Associate Dean of the College of Semiconductor Research at NTHU and previously held the position of Dean of the College of Engineering.