I think to be a nimble startup starts with having an idea that is flexible enough to be adjusted as conditions change.
So, if you are designing an electric car, and you are going to use a battery from company A, and you have everything so tight that only that battery will work, what happens if there is a better batter that is cylindrical, rather than having flat sides? Can your engineers quickly change the design? Or, have you designed with enough space that you can move components for a change? Basically, how flexible is your design and design process?
You need to know where your inflexible points are. For example, if you decide to write software for .NET, then if MS suddenly changed their prices, you are stuck, as changing to a different platform may be too costly.
By knowing where you can be flexible and where you cannot, you can then focus on tracking changes where you are flexible, but, that also depends on where you are in the design process.
If you have started going through testing on your car, as you have a prototype, then changing the battery may be a bad idea, that may be for version 2.
That is the other thing about being nimble, accepting that you need to ship at some point, and just start to plan on a version two, but, try to keep the interface or controls similar enough so that users don't have to relearn how to use your product. For example, if you used drive by wire in your car, and used a joystick, that may be a good design, but we want cars with steering wheels, as that is what we have experience with, using a joystick will lead to slower acceptance. So, even if you decide to use drive by wire, and later change to fiber optics (drive by light) then as long as you are still using a steering wheel the user won't notice a difference in how to use it, but they may notice the car being more nimble.