Magnetoresistive random-access memory (MRAM), as a promising non-volatile memory technology, has attracted extensive research interest owing to its unique combination of high operating speed, exceptional endurance, low standby power consumption, and CMOS process compatibility. In this review, we provide a comprehensive overview of the technological evolution of MRAM, spanning from Toggle-MRAM to spin-transfer torque (STT)-MRAM and then to spin–orbit torque (SOT)-MRAM. The working mechanisms, performance trade-offs, and integration potential of each generation are systematically summarized. Furthermore, the diverse applications of MRAM—including embedded systems-on-chip (SoCs), edge computing, aerospace and automotive electronics, artificial intelligence accelerators, neuromorphic computing, and hardware-level security—are thoroughly discussed. We also identify key challenges hindering large-scale commercialization, such as the trade-off between write energy and speed, process complexity, storage density constraints, and cost competitiveness. Finally, emerging research directions are proposed, emphasizing short-term priorities such as write current reduction and yield improvement, as well as long-term development strategies focusing on material–device–algorithm co-optimization and ecosystem establishment.
Wang et al. (Wed,) studied this question.