The drive to become self-sufficient in regards to semiconductors has crossed the pond, with European Commission President Ursula von der Leyen announcing her intent to pass a European Chips Act in Davos, Switzerland this past September. Her announcement follows the similar announcement and subsequent passing of the USA CHIPS Act, a $50 billion investment in U.S. domestic semiconductor manufacturing.
A Look At The Act
The act, described by Von der Leyen as a matter of “tech sovereignty,” aims to help Europe quadruple semiconductor production and achieve 20% market share for EU-made chips by 2030. Experts estimate Europe will earmark around $160 billion from its Covid-19 recovery fund to pay for the admittedly ambitious effort. It
“Yes, this is a daunting task. And I know that some claim it cannot be done. But they said the same thing about Galileo 20 years ago,” Von der Leyen said, referencing the EU’s successful decision to design and deploy its own global satellite navigation system for strategic purposes.
“And look what happened. We got our act together. Today European satellites provide the navigation system for more than 2 billion smartphones worldwide. We are world leaders. So let’s be bold again, this time with semiconductors.”
A final version of the act is expected to go before EU member nations in February, but von der Leyen outlined a five-pointed approach to innovating Europe’s semiconductor industry:
- Strengthen research and innovation capacity in Europe largely through support for the Alliance on Processors and Semiconductor technologies, a part of the European Commission formed in July 2021.
- Focus on ensuring European leadership in semiconductor design and manufacturing.
- Adapt state aid rules to prevent grants and policies at the nation-state level from undercutting a collaborative EU effort and to allow for public funding for European production facilities.
- Improve Europe’s ability to anticipate and respond to shortages in the market. Always a great idea.
- Invest in and even take equity stakes in high-tech startups via the European Innovation Council, Europe’s program to scale up breakthrough technologies on The Continent.
The act comes as German chip manufacturer Infineon opened a $1.9 bn plant in Villanch, Austria, Intel eyes a $95bn investment in Europe–most likely at their existing high-tech Leixlip, Ireland facility, and fab giant Taiwan Semiconductor Manufacturing Company (TSMC) entered the earliest of talks with the German Government to build a facility there. Intel, however, just announced big-but-vague plans to build a huge $20bn (or is it $100bn?) facility in New Albany, Ohio.
This means their decision as to exactly where they’ll invest and how much will probably come after the details of the EU Chips Act become public. Building a new fab is a risky proposition–it takes time, an outrageous sum of money, and the chips could be old by the time they roll off the line. Intel and TSMC will look for help in the form of subsidies and grants to offset these risks.
The Response? Mixed At Best
As one might expect, government promises of highly innovative things to come and a corresponding need to urgently spend lots of money were met with suspicion. Many, many writers voiced their apprehension, but this article from The Bangkok Post does a great job of outlining the potential pitfalls of the EU–or any government–expecting capital to work miracles in the semiconductor manufacturing arena.
Most arguments for domestic semiconductor manufacturing rely on the idea that semiconductors are vital for national defense and Asian-dominance in their manufacturing puts the EU, or the US, at risk, especially should tensions with China become open conflict.
Before we get into the geopolitics of it all, let’s take a step back and take a look at semiconductor manufacturing for a moment. It’s a hugely complex, global process that requires very specific raw materials only found in certain places on earth and consists of three types of players:
- Designers like Nvidia that do R&D, create the software and intellectual property behind a chip but don’t make it. Companies that design but don’t fabricate chips are called fabless.
- Foundries like TSMC that manufacture chips designed by fabless companies.
- Outsourced testing and packaging for delivery to electronics manufacturers.
It’s in the foundries that the only real “Asian dominance” exists. 80% of the chips made today come from Taiwan and South Korea. But that’s not where the most important part of the semiconductor supply chain is happening.
Design accounts for almost half of the value of the semiconductor market and that’s still a North American dominated market. All three top suppliers of sophisticated electronic design automation software (EDA), vital for chip design, are American. If you need an extreme ultraviolet scanner, and you will need one if you want to draw circuitry on the most advanced chips, the only place to get one is ASML in The Netherlands.
And as for the threat from China, this op-ed (and others) posit that China is at least 10 years behind the curve in technology, and with the U.S. refusing to sell advanced chips and chipmaking tools to Semiconductor Manufacturing International Corporation, China’s largest domestic (and partially state owned) fab, it stands to fall farther back in the global race to chip dominance.
China, though, happens to be one of the world’s largest importers of semiconductors, to the tune of $378 bn in 2020, according to a whitepaper by the Semiconductor Industry Association, a US based trade group. If firms lose access to Chinese sales as a result of the political ramifications the results could be disastrous.
The U.S Chamber of Commerce estimates that decoupling from China could cost the U.S. market between 8%-18% of its global market share and the loss of 124,000 jobs as a result.
All of this brings to a head one of the largest points made in The Bangkok Post and other critical op-eds is that the government or industrial policy isn’t the best tool to solve the current shortages.
All of the proposed new fabs are years away from completion, and there are already whispers of a semiconductor recession in Q4 2022 or Q1 2023. These, and other issues like ever changing consumer preferences, labor costs, and the rapid pace of development, are best left to the market to solve. Our piece on the US CHIPS Act takes a look at SEMATECH, $500 million industry consortium brought together to battle the supposedly inevitable Japanese takeover of the semiconductor industry. Nobody can tell if it worked.
Only Time Will Tell
As we’ve mentioned before, semiconductor plant construction is a long game. There will be plenty more market upheavals and disruptions before any of the new plants come online, and there will be plenty more after. There will be other solutions and innovations that crop up during the time that the EU gets up and running. Count on it.
One would hope that the EU Chips Act does all that it promises. You’d hope that smart strategy and international alignment and cooperation would win out.. That’s certainly our hope at Supplyframe, but only time will tell.