Ensemble several instances of same scikit-learn regressors on different parts of the dataset

Question

I am currently fitting an sklearn regressor on a very large dataset (think billions of rows), and there is no way that I can run that regressor on the whole data without getting