I'm curious what definition of "addiction" you're using to arrive at the conclusion that caffeine isn't addictive.
There's no one standard definition of "addiction", and most concise definitions (one or two lines) are extremely fragile and break down when you try to apply them in any meaningful context.
By this definition you're putting forth, caffeine certainly could be addictive, as could any substance - and in fact, nearly any behavior, which is why professionals generally frown upon using this measure definitionally, because it leads to pseudoscientific terms like "X addiction", for absurd values of X.
You may also be interested in studies on caffeine dependence, like: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3777290/
> Numerous controlled laboratory investigations reviewed in this article show that caffeine produces behavioral and physiological effects similar to other drugs of dependence. Moreover, several recent clinical studies indicate that caffeine dependence is a clinically meaningful disorder that affects a nontrivial proportion of caffeine users.