When is the Graph Neutral Network inductive?

As I have learned, the statement is that “GNNs are naturally inductive because they learn the same neural networks on all the nodes and edges.” Does it mean that all the GNN frame like GCN and GAT can do inductive representation learning?
Also I saw a blog where the writer said that he chose the GraphSage because it can achieve inductive representation learning.
Does it mean only the GraphSage can do this?
So, am I correct in my thinking? Thank you so much!