Is there any way to use Additional regressors without including it in prediction dataframe?

As I understand, We have to use Additional regressors values in both fitting and prediction dataframes. What if I don't have regressor values for prediction dataframes.
Ex- If I'm building a Bitcoin prediction modal, I will choose "Date" and "Adj Close" as my two inputs. What if I want to "Open" as an Additional regressors? There will be no data for prediction dataframes.

Is there anyway I can forecast using three inputs.
Thanks.