Running model on parallel on google cloud

can i run multi-model on the same server trained on different data set, in order to manage the cost ?