On the other, I still think it's worth the time. Being forced to write big, gnarly programs (like a compiler or collaborative text editor or puzzle AI) will give you experience that you'll never get from tutorials online about dealing with big, scary codebases. Or working in teams. Or whatever else.
If you learn as you go, each of those problems will get you fired when you hit it, because they're not exactly intuitive to deal with and you won't notice it until shit has already hit the fan.
There's not a 1:1 mapping between academic and real world problems, but I guess I'm saying it's still going to help. It also puts some extra items in your toolbox in case you want to do something other than vanilla web dev.
That being said, I have no way of evaluating the financial or chronological opportunity costs, as those are unique to you. It also depends how keen you are to get involved w/ your professors and students.
Plus, universities are where co-founders live!