This requires space for two files, the video file and the audio file. After merge, which is a separate step from download (the "dl" in yt-dlp), the extra file is deleted. Whereas with itag 18, only space for one file is needed and no extra merge step using ffmpeg is required.

yt-dlp can't do this without ffmpeg

You need FFmpeg installed, yes. But it's not a separate step you have to do.