Because the platform you create is not a DFX platform, PL kernels are implemented and packaged with platform all the times. They are included in BOOT.BIN. They are downloaded to hardware during boot time.
When the host application is launched, it reads the kernel metadata from xclbin file to understand the kernel topology and parameters. This step will not configure PL bitstream.
AI Engine has two run methods: configure during boot or load graph in host application.