Another thing to consider is that the large projects mentioned may have been built up from a series of small projects you haven’t heard about. So to you, they seem like large, unproven ventures that came out of the blue, but they may have been incrementally funded and marketed over a series of years. Each time the project is successful, it gets more funding to proceed, and whoever signs off on it then has a vested interest in it succeeding. The bigger it gets, the more jobs are involved and the more reputations at stake, which tends to make for a self perpetuating machine.
For example, in the defense field, there is something called the Small Business Innovation Research (SBIR) program, which also exists in the DoE, DHS, NIH, NASA, and other agencies. Speaking from personal experience doing these types of grants, the agency puts out a wishlist of things they would like developed. Sometimes they are very specific in nature, but some agencies, like the Defense Advanced Research Projects Agency (DARPA) tend to be more “big picture” in nature. The way the program works is that you and several other small businesses each get $150,000 to do a proof of concept and show the solution is feasibly possible within 6-9 months. Whoever the Government deems to have the best feasible solution then gets invited to provide a proposal for between $750,000 - $1M to build a prototype. From there, the “sky’s the limit”. If you’ve built an interest technology that has applications across a number of industries and fields, you can get tens or even hundreds of millions of dollars. As part of your marketing efforts, you have usually already started pitching this claim early on.
For example, if I have a new way of combining information from some sensors for an esoteric application to detect, say, when aviation fuel has become contaminated in the field, I’m going to say that it has applications for EVERY fuel pump in the commercial world. That will save $X trillion dollars in terms of repairs to vehicles, reduced pollution, etc. Then I say it will also help with determining when ANY fluid is contaminated beyond fuels, including groundwater, medicines, etc. adding an additional $X trillion. Then I’m going to cite how this is going to extend the life of some other system like hydraulics because the fluid will only get changed when it needs to rather than after X amount of months of use, which is going to reduce downtime of manufacturing machines, which will save $X trillion as well. You get the idea. So by the time we get this far down the road, I think the idea is presented to Congress as “sure it will cost a billion to build, but it will create 50,000 jobs in your district, and it will ultimately save $X billion more down the road and make you look like a hero for funding it”