The way Git thinks of data is different from the way many other version control systems think of data. Many version control systems store data as a list of files and changes to the files over a period of time. They are delta based.

Git on the other hand thinks of data as a series of snapshots. So, whenever you make a commit, Git takes a snapshot of the entire state of project at that particular point of time and stores it. It also maintains a reference to that snapshot.

So let's say at the start, you have 3 files: A, B and C. When you make a commit, Git takes a snapshot of the entire state of project at this point and stores it.

Now, if you edit files A and C and make another commit, Git takes another snapshot of the entire state of the project at this point of time and stores it. Since file B has not changed, Git doesn’t store that file again. It maintains a link to the previous identical file it has already stored.