Null pointers have nothing to do with the instruction set architecture, except as far as they are often represented by the value 0. Can you describe the scheme

Null pointers have nothing to do with the instruction set architecture, except as far as they are often represented by the value 0. Can you describe the scheme you're imagining, whereby their use saves transistors?