There are a lot of complications with switching stacks, but they are not insurmountable. Here's what I think (my day job is .NET, everything else I do is open source).
Cost
This is why most people want to switch to open source, or at least away from .NET. It almost always costs more to set up a decent server infrastructure to host .NET. If you go with a standard cloud infrastructure, most hosting companies will charge between 3 and 10 times as much per month for the basic plans.
Where it gets tricky is measuring long-term costs of development on a stack you don't yet know vs. opportunity cost of developing on one you do. However, if you're like a lot of bootstrappers, you don't have a lot of money to spend on server hosting, and it has to be as cheap as possible.
As an example, I spent approximately $150 my first year in hosting and bandwidth charges for 1 small server while I was working on setting up my business. If I had chosen to go with a Microsoft server, it would have cost more because they simply don't allow as small of servers. And, SQL Server is super expensive. It would have cost almost $2000. A comparably powered Ubuntu server would have cost around $500.
You
A lot of it depends on your ability to adapt and learn a new stack. It's not easy, because as a startup it's about much more than just learning a programming language. You have to set up the servers, manage the infrastructure, the databases, everything. There are services out there now that offer Platform as a Service and can abstract a lot of that management away for you, but that is often more money. You'll have to determine if that trade-off is worth it.
Knowledge Transfer
You say that you hope to adopt the best practices you've learned from .NET and transfer them to a different stack. My experience is that other than very high-level tactics, you're going to pretty much throw out whatever it is you know about .NET web dev when you move to anything else. There's just too many different pieces, different cultures, and different approaches to really achieve that much reuse.
That's not a bad thing though. You really should go learn other stacks, as they will challenge whatever you currently know and make you a better programmer.
Summary
My opinion is that if you're willing to trade off a slower start for a stack that's easier to manage and scale, and at a lower price, you should go with an alternate stack. Other people may say that software costs are not really the big cost, but unless you have funding, a lot of money saved up, or are willing to borrow a lot of money, that's just really not true at all. It's expensive to buy all that software and server infrastructure up front, and if you don't have the money, you can't pay for it.