The one piece of this that might end up costing a ton of money is the testing using live soldiers. Each of those deployments, along with all the heavy equipment is a lot of money. I saw they already have 80k hours of soldier feedback, so in part, that's where some of the money goes. The rest, of course, is figuring out how to make the HoloLens light and reliable, which no one has ever done. Idk, this could be a huge money pit. The Army has tried this idea before and failed. Now might be the time, or this could be another JSF spending fiasco. Either way, MSFT is doing well.