Pretty much yes. In simple terms, when you send instructions to the graphics card you essentially say “Please render this group of polygons/vertices with this shader program”. When you want to draw another chunk, it becomes a separate draw call. This is why you have techniques like texture atlasing where multiple textures are combined in one texture sheet, because it allows several meshes to be combined and use the same shader program.
I haven’t dived into the code and GPU instructions, but I assume the same as Jenny Gore is saying, which is that each chunk is sent together with only the influencing transforms and the related shader program. The GPU only has access to the relevant transforms, which is the limit of 75.