Aart de Geus
Analyst · Rosenblatt.
Sure. Well, let me make 3 comments. First is there have been a number of generations of semiconductor transistors, [ planar ] was for many years the way to do it. The thought to do FinFET was heresy because these were very difficult to do and impossible and then they became possible, and we moved forward on that. The way to look at data all around is essentially, it's a sophisticated variation on FinFET and a number of advanced customers and manufacturing partners are clearly investing in this and clearly going in this direction. And we can absolutely support it all the way from the very beginnings of providing them with tools that allow them to literally simulate and design individual transistors all the way to our design system. Fusion Compiler is completely capable of handling this. So that's not going to be a problem.
The second comment is what you're referring to when you say new types of transistor is de facto a form of continuation of Moore's Law. Meaning still smaller transistors, still less power, still faster speed. And while Moore's Law has slowed down, I want to emphasize, it is still continuing. And that, too, for many times was viewed as impossible, but it is absolutely continuing.
I would amend it with something, though, which is you also see more talk of now bringing multiple chiplets or sometimes called tiles closely together, sometimes on top of other chips, on other pieces of semiconductor. And to me, that is a mega booster pack to Moore's Law because if you cannot do more transistors on a chip, can you do 2 chips closely together? And there are a number of design issues that Synopsys is particularly well equipped for because when you do something in 2 chips instead of 1, the biggest problem is the in between, right, which is how quickly can you get the signals from one to the other. And that's why you want to have them very, very closely linked.
We have a complete 3D IC, that design system for that. And as we see people do a combination of still smaller devices, but now also multiple chips closely together, we will gradually see that growing. So I'm very encouraged that for the next 10 years, we're still going to see way more complex systems. And in our case, the word complex is a good word.