中文 English

A Sputnik Moment For Chips

New funding recognizes supply chain risks, but it raises a lot of challenges and unanswered questions.

popularity

Chip shortages are the new Sputnik moment, and they have created a sense of national and regional panic not seen since the days of the Cold War. For both the United States and Europe, those shortages have sparked some of the largest technology investments by government in the past half-century that are not strictly for the military — and by far the biggest involving semiconductors.

While there are lots of statements of direction in legislation, there is much that needs to happen to make this work. Legislative groups have the right idea that action is warranted to maintain competitiveness, but their understanding about the inner workings of the supply chain and the various pieces of technology required to tighten that supply chain are about as finely tuned as throwing a blanket on the beach. It covers everything, but there are some pretty significant gaps underneath.

In the U.S., the CHIPS Act will infuse $39 billion into domestic manufacturing capability, including $2 billion for legacy chip production, and another $11 billion for workforce development and R&D. There are other programs infused in there, as well, bringing the total number up as much as another $4 billion (depending on who’s counting and what they’re including) for everything from on-shore, university-based prototyping, lab-to-fab technology, as well as advanced security technology. The bill also provides $1.5 billion for public wireless supply chain innovation to spur open-architecture, software-based wireless technologies, and a 25% tax credit for investments in semiconductors and “specialized tooling equipment.”

What’s clear from reading this bill’s vague and confusing allocations, is the latest supply chain glitches have finally grabbed the attention of politicians everywhere. At present, only about 12% of the world’s chip manufacturing is done in the U.S., compared to 37% in 1990, according to a 2020 Semiconductor Industry Association report. In Europe, the European Chips Act aims at boosting European chip production from 8% to 20%, which is what it was in 2000, according to SEMI.

There is no question that having multiple sources for every electronic component improves the robustness of the supply chain. Debate has been smoldering beneath this topic since offshoring really began picking up steam in the 1980s. Up to, and even beyond the turn of the Millennium, economists argued that on-shore manufacturing was environmentally unfriendly. After that the argument was weighted heavily toward labor arbitrage, which was supposed to be good for consumers. These always have been Swiss-cheese-like arguments. With some investments, industry can be made much more environmentally friendly. And outsourcing of high-paying manufacturing jobs that produce tangible goods is never a positive thing. Still, it took nearly 40 years for those problems to warrant significant legislative action in the U.S. and Europe.

So what’s different today? Part of answer is the technology itself. Since the late 1990s, software has been seen as the future of everything. That all worked as long as the cost per transistor dropped in half every two years. But once scaling reached the finFET generation, software suddenly began losing its luster. It’s slow, energy-inefficient, and it requires an almost endless stream of security and performance updates.

This has moved semiconductor technology to the forefront of nearly everyone’s strategy over the past half-decade — even companies like Google, Meta, and Apple (which once went to war with Microsoft over the look and feel of the graphical user interface) are now designing their own silicon. And not just any silicon. These chips are increasingly customized for specific applications, data types, and use models. So there are more custom chips required, more chips produced more slowly in smaller batches, and more demand for everything. And the resulting shortages are sparking concerns (and in many cases, “concern” is a gross understatement) in industries such as defense, automotive, aerospace, household appliances, public utilities and services, and just about anything with a plug and battery. Even toasters are getting smart enough not to burn toast, providing you can find the right chips.

Shortages have allowed companies to gouge consumers, which contributes to inflation. They also are responsible for widespread chip substitutions of questionable reliability and security. And they have forced companies to look at semiconductors as a critical component of national security, and one where China has been making massive investments in everything from chip manufacturing to securing mining rights for raw materials such as rare earths, lithium, and wafer technology.

The dual chips acts are aimed at making the U.S. and Europe more competitive through massive infusions of capital. But redundancy is costly, and government investments — which are basically subsidies — are being made under the banner of national sovereignty and defense. That may bolster indigenous companies in the short term against lower-priced international competition, yet the challenge will be maintaining the same level of innovation and competitiveness for which the chip industry is a recognized global leader. That means government involvement and regulation, particularly with an emphasis on open standards and reusability, will require serious oversight and management, something the chip industry hasn’t seen since its very early days.

The question now is whether governments are up to the task, and whether startups and would-be competitors will be able to penetrate markets with the same speed and adaptability if those investments aren’t carefully managed. Handouts are always attractive to business leaders, and in the case of offshoring, the pendulum certainly needs to swing back for economic and security reasons. But it’s entirely too easy to get comfortable with infusions of free money. It can reduce competitiveness in the long term rather than improve it, and it can have enduring impacts that are contrary to the stated goal of all of these efforts.

The chip industry has become critical to nearly every industry, and these investments are a recognition of just how far this industry has come. But it would be wise to recognize the potential downsides early on, and for the industry itself to manage growth and ensure continued competitiveness from the inside.



5 comments

Raye E. Ward says:

This is a wise piece. Previous efforts have increased industry collaboration; hopefully, the lobbying efforts that have gone into the CHIPS bill’s passage will continue to open standards, reuse and practical innovation. Thank you for sharing your thoughts.

Robert Pearson says:

Key statement – “will require serious oversight and management, something the chip industry hasn’t seen since its very early days.”

This is where the in-fighting will drag us down.

Mike Sottak says:

Hi Ed – thanks for a very thought-provoking and insightful analysis

Gary Dagastine says:

“But it’s entirely too easy to get comfortable with infusions of free money. It can reduce competitiveness in the long term rather than improve it,” — Right on.

Dr. F says:

At present the just passed $52 billion CHIPS bill does not contain any requirements to prevent the siphoning off of new technologies to unfriendly or vulnerable foreign entities by recipients of these funds through channels common in the U.S. semiconductor manufacturing industry (Fabs/Foundries, as well as their suppliers, including EDA). It should. Also, when either the government and the supposed beneficiaries of CHIPS (e.g. the U.S. auto & defense industries) buy systems that contain chips fabbed or packaged offshore, they must demand a solid technical and economic explanation from those suppliers as to why were they not acquired stateside. Old excuses of labor shortages, high labor costs, and unions do not wash — not even for advanced packaging. As long as 30 yrs ago, we had developed at major IDMs in AZ the technologies (flip-chip and substrates, the current workhorses of the industry, worth about $30 billion of production a year, as well as the robotics needed for precision assembly made affordable through the application of advanced algorithms and mathematics), and started fabs right here in AZ. Thanks to those foundational technologies developed, as well as put into high-volume manufacturing in the U.S., at least for advanced packaging (which requires continuous R&D in TMP as well as software support for the robots), even in Arizona the labor cost is only 10% to 13% of total cost of manufacturing.

Leave a Reply


(Note: This name will be displayed publicly)