> Even if the US was totally committed, and decided to build a brand new reactor right now, how long would it take? A decade?
Totally committed? You sure? Ok.
The US could bring at least two dozen new nuclear power plants online per year. It could do that with only modest impact on its massive economy.
National security declaration (probably numerous). Bypass all environmental building concerns and impact reviews. Pull most of the US military out of Europe and all of it out of the Middle East. Mobilize at least 100,000 soldiers in support of the mission (ideally twice that number; clearing land, hauling materials, dealing with protestors, providing area security, restricting access including by drones). Whatever needs to be done, will be done. Redirect several agencies (DARPA, DOE, etc) to almost exclusively working on this mission for ten years. Have Congress declare a draft on the back of the national security declarations, pull relevant scientists and engineers out of the private economy via draft. Suspend all emigration. Federalize the control of all national guard troops, draft them all formally under the command of the US Military and remove them from their states to avoid potential state resistance (numerous states would attempt to resist this very strong-armed Federal plan). Nuclear plants would be reasonably distributed across the country, some states would attempt to resist any plants being built inside their borders, which is another reason you remove their national guard troops. This is serious, so we'll do what is necessary. It's life or death after all, right? If the Supreme Court attempts to block it, expand the size of the Supreme Court and stuff the bench until your votes win.
Use eminent domain where necessary. Limit taking private land as much as reasonably possible; doing so just gives the resistance groups an issue to play up. Fortunately in this case the Federal Government owns a ridiculous amount of land (particularly in the western states).
The US has an epic scale labor force, and millions of healthy people doing nothing that could be drafted into various basic positions in support of this large national project. Jobs! (there's that sound bite) Another plus: absolute secrecy would not be particularly paramount, unlike with the Manhattan Project. You wouldn't need a bunch of Oak Ridges. There isn't anything that special about a modern nuclear power plant, and the US wouldn't be using any radically new technology in these plants anyway.
Initially getting up to speed would take several years (standardizing and getting new workers enough experience, etc), after which you could get to a rate of bringing a new plant or two online in a given month.
Fund it with a trivial Federal Reserve Nuclear QE program. $20-$30 billion per month in 'printing.' The dollar wouldn't even notice, and it'd keep the budget from getting further stretched (which would keep the general population more calm than if you spiked their taxes or took from their social programs).
Task the major banks - all of which are realistically under direct supervisory control of the Federal Reserve and Treasury these days - with helping to keep the US financial system stable (there might initially be some panic), including where interest rates are concerned. The US has the most powerful banking and financial system by a considerable margin, it would be told to accomodate and assist, or else (look what they did for funding the shale build-out, a spigot of hundreds of billions of dollars; private loans involved in the program would largely be back-stopped by the Fed; thousands of companies in the private economy would need loans to ramp up early on; the banks making some money on this would remove a potential resistance point). Put the talking heads on TV to work pushing the agenda favorably, to calm the population; spin how positive this will be for the economy, jobs jobs jobs (just keep repeating that). Do the exact same thing for the business & financial media to soothe Wall Street, talk up how good it'll be for corporate profits, manufacturing, construction, et al. Throw billions of printed magic dollars at the military industrial complex and their partners, to remove (bribe) a potential adversary to the plan. There's a line for everyone: jobs, green, do it for the children, manufacturing & blue collar, energy independence, tell them how great it will make America, etc.
By necessity we'd build and improve a lot of grid along the way as well. There would be several hundred billion dollars spent on that over more than a decade.
15-20 total years for the program (a dozen or so prime building years in there, during which most of the plants would come online). Once the early years of hard work of ramping up, standardizing and getting it all rolling was done, you could relax some of the initial political policies that it took to get underway. The program would also include a considerable investment into the aging 60 plant fleet the US has now. In the end, this would wipe out most of the US fossil fuel economy. It would cost several trillion 2019 dollars, spread out over two decades. The ~20 year program span would give us time to gradually retrain labor away from the fossil fuel economy. As the plants are brought online, you'd start by eliminating existing coal plants and coal jobs first (there aren't actually that many workers in coal these days anyway), with the displaced workers getting jobs and job training to participate in the new nuclear power economy.
How committed do you want to get? How dire is the situation?