96-Kb Voltage-Controlled-Magnetic-Anisotropy MRAM for In-Situ Reservoir Computing with High Endurance ($\geq 10^{12}$), Sub-Ns Operation (0.3 Ns) and Ultralow Power Consumption (40 fJ)

Voltage-Controlled Magnetic Anisotropy (VCMA) Magnetic Random Access Memory (MRAM) offers ultrafast operation with high endurance, making it an ideal platform for edge computing applications such as reservoir computing (RC). We propose a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$96-\text{Kb}$</tex> VCMA-MRAM-based RC system for temporal data processing, leveraging sub-nanosecond switching (0.3 ns), ultra-high endurance (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&gt;10^{12}$</tex> cycles), and in-situ training. The RC system demonstrates high recognition accuracy (97.8%) for tasks like voice recognition while achieving energy efficiency (40 fJ per input). These results establish a Si-platform-friendly architecture for scalable and efficient RC implementations.