Paul van Gerven

29 September

ASML CTO Martin van den Brink dropped a bombshell recently: the transistor could stop shrinking in ten years or so. The implications are far more serious for ASML itself than for semiconductor technology.

Martin van den Brink already predicted in 2007 (link in Dutch) that it will be economics, and not physics, that’s going to end chip scaling. Now, fifteen years later, he suggests that the end of the road may be in sight. Although ASML is looking into a successor of high-NA lithography, the CTO seems far from convinced that hyper-NA will prove economically viable. It would take a handful of technological tours de force to make it happen, he intimated.

The truly paranoid might consider this interview part of some sort of high-stakes poker game about who’s going to pay for the development of hyper-NA. I’ll admit to considering that option, but only very briefly. It’s pretty clear that EUV optics are getting prohibitively large. A redesign is required to split big ‘monolithic’ mirrors into many smaller units that can be manufactured in serial processes – if not for the economics, then for the practical matter of transport. Are there even planes large enough?

Besides, a company in the position of ASML doesn’t need to play games. If it deems hyper-NA too risky and only wants to proceed if customers have more skin in the game, simply asking will do just fine. It’s not even a given that chipmakers will be interested. Transistor shrink has been slowing down and at some point, scaling won’t make sense anymore.

Piggybacking

Does that imply the demise of Moore’s law? No, because even though Gordon Moore’s famous observation-turned-self-fulfilling-prophecy has been synonymous with scaling for decades, there’s more to advancing semiconductor technology than the reduction of component dimensions. Device-level innovations, design optimization, advanced packaging and software tweaks will continue to provide power and performance gains, if not on the device level then on the system level. TSMC chair Mark Liu last year predicted that system performance and energy efficiency will continue to advance at the historical rate until 2040 despite transistor-level performance flattening out.

Reports of Moore’s law’s death will nevertheless continue to pop up. They’re greatly exaggerated, often for some self-serving purpose. For example, when Nvidia CEO Jensen Huang once again pronounced Moore’s law dead just last week, it’s because this fits his narrative that GPUs advance at a much faster pace than traditional CPUs, suggesting that the former don’t need Moore’s law while the latter are hopelessly dependent on it.

In reality, GPUs have been piggybacking on Moore’s law just like all other leading-edge semiconductor devices, but Nvidia has also been adding clever innovations in design, software and AI – exactly the kind of system-level scaling we’ll see a lot more of the coming decade.

Philosophy

So, Van den Brink’s comments don’t point to a day of reckoning in semiconductor technology. But what about ASML itself? The prospect of the lithography roadmap ending is daunting. Without shrink, ASML will eventually cease to be an innovation-driven company and turn into a primarily cost-driven enterprise. It would become a mere shadow of its former self.

Former CEO Eric Meurice once told me that ASML’s raison d’être is lithography. Despite diversifying into metrology and other yield-enhancing technologies over the past decade, it’s fair to say that the company has stuck to that philosophy. Not that it had much choice in the matter. Every pair of hands was needed for the core business.

In the next ten years, however, I suspect we’ll see more diversification from ASML. It might not even shy away from venturing beyond its main focus of front-end lithography. Some of the most exciting future innovations in advanced semiconductor technology will be seen in the back-end. Perhaps that’s a good place to start looking.