We are AISing. We developing and licensing new types of Edge AI algorithms based on tree data structure. We have published two algorithms and will add another soon. All the algorithms are capable of making prediction, with incremental learning if needed, in microseconds cycle time without Internet. The memory requirements are much more compact compared to deep learning, ranging from Kbytes to 10Mbytes to cover time-series signals on a single-threaded, embedded CPU without GPU. These algorithms can be integrated into customer’s control system and improve its optimization and prediction accuracy drastically.