Honestly, at that level it kind of isn't. If you don't agree, you probably also think that high schoolers would benefit from starting businesses, which suggests more that you're a businessman among engineers than that engineers should be learning to run businesses. There are two sides to a successful startup, business and technical, and post-bubble you probably don't have the technical side fulfilled as an undergrad. Failing because you have a half baked implementation looks a lot like failing because you didn't find the right customers, so now you learned the wrong lesson.
The second time around, you'll make damn sure there are customers before you build.
This means your first startup was wasted because you weren't even thinking about the most basic aspects. You didn't learn anything other than that you started off clueless, which a mentor would have saved you from.
I'd argue it's a very valuable experience, and you even if/when you fail with that first startup, you come out with a tonne of skills and valuable insight.
You'll learn a lot from a failed startup, but I don't think that you'll learn as much as if you worked for a startup instead. Learning by direct experiment is less efficient than learning from someone else's mistakes, which is the whole reason you're at a university to begin with.
You would probably be better served by a weekly study of actual startups - how they began, what they created, how they ended, and invite local founders (business and technical) to speak. Combine that with a non-profit technical project that delivers a product and tracks adoption but avoids the requirements of profitable business and you'll impress the community while building a résumé that will help you raise VC when the time comes.
Anyone can throw pasta at the wall; Tech can do better than that.