Future of Moore's Law (and what it means to investors)

“Moore’s Law is dead.”

- Jensen Huang (CEO, Nvidia Corporation), September 2022

"Moore's Law is alive and well"

- Pat Gelsinger (CEO, Intel Corporations), September 2022

Moore's law refers to doubling of number of transistors in integrated circuits every two years. It is the underlying force that has been driving most of the innovation that we have seen in our lifetime. Increasing internet speed, powerful smartphone & notebooks, your Instagram feeds, all of them are driven by this powerful force called Moore's law.

People have been speculating for decades that end of Moore's law is near. Gordon Moore, who postulated the law himself speculated in early 2000s that Moore's law will be dead by 2025. His speculation was based on the fact that transistors ultimately have to be larger than the size of the atoms. A silicon atom occupies a space of around 0.5 nm whereas the current best in class chip from a TSMC or Samsung foundry has transistors of size of 5 nm. I am not the one who knows what the final limit will be. But we are fast approaching that limit. Of course players like Intel are claiming new approaches like 3D architecture and using different elements will circumvent the problem. Maybe they will be successful. But I take Intel's claims with a pinch of salt given that they trail Samsung and TSMC in terms of minimum feature size. Their first chip using the fancied 3D architecture (Lakefield) was withdrawn barely after 1 year of introduction. And their next chip with this architecture (Meteor Lake) is reportedly only slightly better than the cutting edge chips currently, not a generational improvement which is needed to keep Moore's law going.

While jury is out in open when we will hit the physical limits of Moore's law, there is a more mundane issue limiting the continuation of Moore's Law, the issue of economics. A whitepaper by Aalbun, an Intellectual property service company, estimates that ?jump in R&D costs from improving a 10nm minimum feature sizes on chips to 7nm minimum feature sizes on chips is $100 million, and then from 7nm to 5nm minimum feature sizes on chips is $550 million. Thus the costs are exponentially increasing even with linear reduction in feature size. The know-how and investments reflect in reduction in number of foundries producing the cutting edge chips. World over only 4 foundries are producing sub 10 nm chips and only 2 have gone to 5 nm so far.

If you look at the economics angle, there is no question that Moore's law is coming to an end. The only question is when. Since I don't know exactly when Moore's law is coming to an end the rest of the passage is more speculative and focuses on what will happen if indeed Moore's law comes to an end.

Implication for Investors

From one perspective, end of Moore's law will be an absolute tragedy. The world of last 50 years is built on the Moore's law. Without Moore's law there would be no internet, no 4G, no smartphone, no PC, no email and no social network. Moore's law has been been the primary driver of innovation in the field of information technology which we today take for granted. Without Moore's law, the wheels of that innovation will slow down. If I have to put it more dramatically, the wheels of the third industrial revolution will come to a halt (though there will be a fourth industrial revolution in future).

However, for many investors it might be a blessing in disguise! While the continuous innovation in information technology field have led to multiple billionaires and many successful investors, it has been done with an overhang of a threat. Except for the lucky one's like Google, Facebook, Amazon, Amazon etc. which had enormous network effects and brilliant founders to their advantage, most other always kept facing threats from a new generation of competitors. With continuous development in technology, a new breed of competitor could build a much better software at a lower cost compared to the incumbent. AOL, Yahoo, MySpace, Palm Pilot, Nokia, and hundreds of other companies faced the consequences of constant march of technology. Change is an enemy of an investor.

With this change slowing down, the certainty for investors will increase. Companies which are incumbents today will likely remain incumbents for much longer compared to their predecessors. If if it was very difficult to dislodge Google or Microsoft from their leadership positions when technology was moving with a fast pace, it will be even more difficult to dislodge them when the pace of technological change becomes slower. A small startup will find it that much harder to create something new which will make an incumbent's piece of technology obsolete. This means the incumbents can keep making money for much longer than was the case earlier. The road to investing in the technology industry with certainty will become that much wider.

However, such a scenario might be at least a decade far in future. The progress of current technologies and further scope of improvements in chips will be able to sustain the momentum of innovation at least for another decade if not more.

The end of Moore's law will have an implication on the chip industry as well. Till today, a TSMC or Samsung was forced to keep investing multiple billion dollars in new foundries before their competitors did, just to be able to remain relevant. TSMC and Samsung became the largest foundries in in the world by being one of the first in investing in 180 nm, 130 nm, 90 nm, 65 nm, 10 nm, 7 nm and 5 nm foundries. But once, there is an end to Moore's law, they would not be forced to invest tens of billions of dollars in trying to reduce the feature size further. Instead that money would be available to be returned to the investor. Till multiple other players figure out how to reduce the feature size to the same level as a TSMC or a Samsung, these incumbents can simply keep churning out cash from their foundries without the need of further reinvestment. And going by how complex modern foundries are, it may take decades for other players to catch up. The titans of chip foundries will become even more formidable and even more attractive investment option.

If you are looking for the wheels of technology to keep turning at the current pace, pray that researchers figure out a way to circumvent the obstacles in front of Moore's law. But even if they can't, you will at least have a method to keep making money for a long time.

Disclaimer

Abhishek Chauhan is Principal Officer of Eklavya Capital Advisors, a SEBI regulated Portfolio Management Service Provider.

The information, opinion, or analysis provided in this article is not an investment advice and is intended for general informational purposes only.?The contents and information in this document may include inaccuracies or typographical errors and all liability with respect to actions taken or not taken based on the contents of the article are hereby expressly disclaimed.?

Any reader or this article should refrain from acting on the basis of this article without first seeking independent advice in that regard. The views expressed in the article by the author is in his individual capacity only.