In yet another of its questionable design decisions, SciKit-Learn chose to end the names of its hyperparameter tuning classes CV (e.g., sklearn.model_selection.GridSearchCV). The SKL hyperparameter search classes do use cross-validation by default (hence the name), but it can be disabled.
To be absolutely clear: hyperparameter search is not a kind of cross-validation, and does not imply cross-validation. It is often a good idea to cross-validate when performing hyperparameter tuning, though not always; consider, for example, a large neural network with plenty of training data.