My main issues with disabled buttons is 1) you are assuming the user knows your color palette, and 2) if people see a button they will invariably click on it.

I have watched hundreds of user testing videos where I see users mashing the disabled button and not understanding why it isn’t working (of course you can add error messaging).

I prefer the pattern of not showing the button at all (in the use case shown) until at least one character is entered into each text box. Then the button can be surfaced and any front end validation can be tied to the button click.