How does SVM behave when data is linearly separable with large margin?

SVM finds a stable hyperplane with zero training error and maximum margin.

When data is perfectly separable, SVM selects the hyperplane maximizing margin, leading to sparse support vectors and strong generalization guarantees.