What Anduril is trying to do is to first build product that they think the customers will need, on their own dime, and with no exact specs from DoD or anyone el

What Anduril is trying to do is to first build product that they think the customers will need, on their own dime, and with no exact specs from DoD or anyone else, and then present the (more or less) finished product to the customer, and tell them to take it or leave it. This removes the need for negotiating R&D contracts, preparing detailed spec or working with someone else’s detailed spec etc, and does cut significant amount of process related red tape. The risk here is that the customer will, indeed, leave it instead of buying, which means that R&D costs are not recouped.