5G NR massive MIMO systems require intensive computation for matrix multiplication. It is shown that 100 MHz 64-antenna 32-layer 5G NR beamforming can be implemented on 64 AI Engines with three kernels with slight differences, while the same functionality needs 2048 DSP58 blocks and hundreds of thousands of LUTs and FFs to be implemented on programmable logic. The proposed matrix multiplication implementation on AI Engines has a flexible and scalable architecture applicable to a wide range of use cases including, but not limited to, 5G wireless.