a 1280*720 frame, 16 bit, ping-pong buffer uses only 3.6MB of ram, still plenty of ram left for compression. The Keyboard/mouse component use like 512KB ram at most(even less, HID is cheap).

That was my first thought too, but it seems like they're using two of them in the wireless version, with a C5 for wifi

The only reason it had been difficult was video capture and compression. Even 32MB would be overkill if you didn't need remote screen.

That board still requires an HDMI module. The pi zero will support PiKVM fine with an HDMI module too.

The ESP32-P4 is a beast

looks nice but seems to be rather expensive