My hope: whichever direction Gimp development goes, they should try doing something similar to Blender - where they bring a bunch of developers and artists into a space. While they're in that space (for a limited amount of time) work on some big project involving Gimp. Develop Gimp to suite that project's needs.
So the broad decisions about what will be done and in what way have already been made at the team- and individual-psychology level, assuming this money doesn't radically change the core team composition. While the objective manifestation of the spending is yet to come, this money was already spent by the team psychology, and the further we depart from GIMP's past in wishing for changes, the less likely our wishes will match with the actual outcome of the expenditures.
And let's take the polar opposite into consideration--even if the idea of becoming more like Blender and gathering people and putting a big project together makes the GIMP team members recoil in horror, the most well-founded hope would be that the GIMP team builds on past successes and learns from past failures, or gains more momentum in the direction that will likely bring it the most success _given the team's psychological makeup & personal resources_. That would probably bring about a really good outcome even if it doesn't result in a highly-visible project.