Sure, OSF costs money, but VC-funded startups are under pressure to monetize their customer bases, whereas grant-funded projects just need to make something useful enough to justify continued investment. There's the obvious risk is that a VC-funded company will make its money off of selling your private information. (I don't think this is very likely, but I trust Brian Nosek a lot more than I trust most startups.) The bigger issue is that VC-funded companies need to keep a tight leash on their product, whereas grant-funded projects are free to grow in a distributed fashion. While ordinary people are happy to live within the limits of their software, scientists like to explore new frontiers, and it sucks when our software can't take us there. This is why I firmly believe that software (and preferably hardware) that scientists use to do science needs to be open source.
I'm not sure I'd call grant income that gets reinvested in infrastructure and educational expenditures "profit." Not only do private corporations buy shiny buildings, but they also have a legal obligation to make as much money for their shareholders as they possibly can. MIT doesn't.