Agile development is really more suited for huge monolithic developments because those are the ones where people actually know least about what they're doing upfront. Creating a technology demonstrator is an example of an agile strategy.
Building a smaller modular ship before building an aircraft carrier would be another example. It's about finding an incremental path. If you try to do otherwise you just end up spending the money anyway and more on all the mistakes that you make but you haven't had a chance to limit the scope and cost of those mistakes.