EUV step and scan systems are roughly the size of a bus, fractal system complexity and sophistication that rivals the LHC, and mirrors with a roughness two orders of magnitude finer than what the JWST space telescope has to work with (namely .2nm compared to 20nm).
Please do not underestimate the difficulty of manufacturing EUV systems, ASML only manages to build a few dozen a year for a reason :)
Just call the three or four companies you think would be the closest and tell them they’ll be competing to build the capacity paid for by the defense department.
There are many example of big government funded projects that crumble under their own weight, despite being inundated with cash. Look at NASA's SLS, the DoD's trillion dollar plane projects, big government software projects in general.
For truly large, complex systems, it's not enough to throw more money at the problem. You need an organization that can manage the complexity in the first place, otherwise you can spend as many years and trillions as you want with little to show for it.
You can't suddenly create experts by throwing money at random business people.
If all it took was a few billion dollars, don’t you think someone would have already funded it and started a profitable company?
That's not to say that is impossible, but without good examples we can all say a bunch of stuff and it won't mean anything.
You can look at the history of the Defense Production Act (1950) to see it in action several times building up local industries with lots of cheap money.
You cannot just throw any bodies at a problem like creating ASML. You have to throw very specific bodies with each one having at least a decade or more of very specific expertise (it seems like, based on the comments and articles I have read). That is just ASML, not even getting into the deep expertise TSMC must need to execute, or even Apple in designing.
Knowledge itself has expanded tremendously since WW2 such that to be top of the field in WW2 required much less time than to be top of the field today.
“””
bachElor’s EnrollmEnt and dEgrEEs Enrollment in undergraduate engineering programs increased this year continuing a decade long trend to 622,502 full-time students in 2018. Degrees awarded increased to 131,937 for reporting institutions (136, 233 including non- reporting institutions). The three engineering disciplines producing the most degrees in 2018 were, as in 2017, mechanical engineering, with 31,936 degrees, computer science (inside engineering) with 19,082 degrees and electrical engineering with 13,767.
mastEr’s EnrollmEnt and dEgrEEs There was an increase in 2018 in enrollment in master’s engineering programs with a total of 93,559 students enrolled, a 6,687 student increase over 2017. The number of master’s degrees awarded increased to 66,340 for 2018. The three disciplines representing 39% of all engineering master’s graduates are computer science (inside engineering), with 10,946 graduates, mechanical engineering with 8,160 graduates and electrical engineering with 7,048 graduates.
doctoral EnrollmEnt and dEgrEEs In 2018 there were 12,156 doctorates awarded 78,715 full-time students enrolled in doctoral programs. The 5 disciplines with a combined total of 60% of the doctoral graduates in 2018 are mechanical, electrical/computer engineering, electrical engineering, chemical engineering and computer science (inside engineering
“””
Fermi, Teller, Bohr, von Neumann, Bethe, Szilard, Einstein - not every one of them was directly involved, but they all moved to America and taught younger generations of American physicists.
If it is, at best only temporarily. The US is perfectly capable of setting up its own complete supply chain if there were the political motivation and do so in a relatively short time frame. China is perfectly aware of this so it will never become a serious issue, they would just be shooting themselves in the foot and cutting themselves off from our (sometimes unwilling) technology transfer.