The boundary that was important at the time was not caused by the number of atoms, but by other laws of nature, such as the diffraction limit (“How thin are the lines in light?”). These laws of nature have not been broken, they have only been worked on. First with submersible lithography (literally: put a drop of water between exposure and foil) and multiple etchings (multiple exposures that together form the desired pattern), but then the fun is over. At that time, the harsh laws of nature made further progress impossible. Well, there was one trick left: use a light of a different wavelength, EUV or something like that. Unfortunately, this is also not possible, because this type of radiation is absorbed by almost everything. Not only through the glass of the lens (so you will suddenly have to work with mirrors), even through the air (sweeping the whole machine is really inconvenient). But wait a minute, these are big problems, but they are “only” financial problems, the laws of nature go along with them. So, yes, it is quite expensive, but if you want it powerful enough, you can do it. So (by heart) Intel, Samsung, and TSMC bought billions worth of ASML stock so ASML could use all that money to do the “impossible” anyway.

Don’t get me wrong, I’m in awe of the engineers at ASML, but let’s face it, as long as transistors are made of atoms, a transistor will never become smaller than an atom. Will we ever have a lithographic process that makes the width and length of transistors a few hundred picolitres? I look at them. But less than a hundred picometers? No sorry, I don’t see how you (even if you have many billions available to search) can make atoms smaller than they are or make transistors smaller than an atom.