Wouldn't scientific publishing count as non-profit? What about open access publishing/preprints/blog posts? There is an increasing grey zone of what constitutes "scientific publishing".

I think this is something that you need to contact Wolfram, Inc. directly rather than through this forum. You probably should be prepared with your license number and an answer to the question about if this scientific publication requiring (or at least using) Mathematica is part of your job duties and/or if the publication leads to enriching your or your colleagues either directly or indirectly.

For example, SAS (a statistical package) has a "University Edition" free to download which on the associated community website has answers to such questions (again, about SAS). In essence, if money is involved (directly or indirectly), the answer is "No".