How does VLSM save addresses if we can just use NAT?

Question

I understand how to configure VLSM, but I'm still confused why it would be needed if we have Network Address Translation. My understanding is that a local area network can use NAT to have a user or a large amount of users with private IP addresses, but still use a single unique public IP address for all those users. So what if 10-242 private addresses are wasted? They are free aren't they? Where does my logic fall apart?

-Artekis

Answer 1

Variable Length Subnet Masks was an effort to enable more efficient use of public address space. NAT enables us to use fewer public addresses, (by enabling usage of private addresses internally.)

So the two ideas aren't directly related, even though both help reduce public address space waste.