As others have pointed out, even a single atom exists in three dimensions. However, there's an alternative, and mathematically rigorous, sense in which you should regard a sheet of graphene (or two sheets of stacked graphene, or four sheets of stacked graphene) as 2D.
In condensed matter, one is often interested in the properties of a material in the thermodynamic limit, i.e. as $N\rightarrow \infty$, where $N$ is the number of atoms in the system. There are many ways of taking the thermodynamic limit. If we consider a cube of material with dimensions $L_x\times L_y \times L_z$, we could take $L_x\rightarrow\infty$ while leaving $L_y$ and $L_z$ unchanged, or we could take $L_x,L_y\rightarrow\infty$ while leaving $L_z$ unchanged, or we could take all three to infinity. We call the first case "one dimensional" no matter how big $L_y$ and $L_z$ were initially, we call the second case "two dimensional" no matter how big $L_z$ was initially, and we call the third case "three dimensional." In other words, a one dimensional system in condensed matter is a system where one direction is infinite, and the other two directions are finite. In a sense it doesn't matter how large the other two directions are, since they are necessarily negligible compared to infinity!
With this definition of dimension, you can prove all sorts of interesting theorems. You can show a continuous symmetry cannot be broken in dimensions $\leq 2$. You can classify topological insulators according to their dimension. And on and on. All of these theorems are proven using the notion of "dimensional" that I defined above.