Centrifuges are delicate but they're not especially difficult for an industrial society to produce. They are comparable to jet engine shafts in terms of both loading and components. The manhattan project didn't utilize them because a single individual overlooked a simple method of making them more stable - you add a flexible section so any vibrations damp themselves out. German scientists claimed to have figured out this trick during the war, and the soviets implemented it shortly thereafter in their nuclear program, allowing them to enrich enough material for their weapons with only a fraction of industrial effort of the American project. Gas centrifuges for enrichment of chlorine isotopes had already been in use since 1934.
Much smaller nation states have since been able to effectively generate enough uranium for nuclear weapons programs in short periods of time using centrifuges. The main challenge for them has not been production itself, but hiding the activity from international scrutiny. Tracking shipments of materials and parts necessary for centrifuge construction has been one of the leading methods of avoiding nuclear proliferation, but still centrifuges are so effective that even a small number is enough. North Korea, for example, is believed to have two buildings worth of centrifuges, but we only know where one of them is, and only because they have given international inspectors a tour of that location. For context, North Korea's current GDP is about 1/64th of the US's GDP in 1939.
"Gernot Zippe, an Austrian physicist, .. in the early 1950s figured out (with others) how to fix the problems that Beams had with his centrifuges. Amazingly, he did this while being a prisoner of war in the Soviet Union."
http://blog.nuclearsecrecy.com/2012/05/28/zippes-centrifuges...
The centrifuge has also come to play an essential role biochemistry. It is used to both separate proteins and determine their properties. A key challenge was to make something spinning above 10 000 revolutions per minute safe.
Edward Pickles took on the challenge:
“... (the Model L) has done more than any other single instrument to advance the study of viruses, yet Pickels received almost no attention for these remarkable achievements. Previously, ultracentrifuges had to be run behind three-foot-thick walls of reinforced concrete. The scientists who operated them were considered slightly insane. Centrifuges required constant attention … and explosions were to be expected at fairly regular intervals. The Spinco preparative centrifuge, for contrast, is about the size of a washing machine, and anyone can learn to run it in 10 minutes.”
https://www.beckman.com/resources/technologies/analytical-ul...
Much smaller nation states have since been able to effectively generate enough uranium for nuclear weapons programs in short periods of time using centrifuges. The main challenge for them has not been production itself, but hiding the activity from international scrutiny. Tracking shipments of materials and parts necessary for centrifuge construction has been one of the leading methods of avoiding nuclear proliferation, but still centrifuges are so effective that even a small number is enough. North Korea, for example, is believed to have two buildings worth of centrifuges, but we only know where one of them is, and only because they have given international inspectors a tour of that location. For context, North Korea's current GDP is about 1/64th of the US's GDP in 1939.