We’ve done a lot of writing about efforts by the US and EU to strengthen their domestic semiconductor production in response to the supposedly looming threat of Chinese dominance in chip manufacturing.
The rationale is as follows:
- Semiconductors are essential for economic innovation and advanced military equipment.
- We (the US or EU) do not fabricate the majority of our own semiconductors domestically.
- China’s ambitious Made In China 2025 plan to increase their domestic semiconductor production therefore represents an existential threat.
Are They Though?
Our elected officials and media added their voices to the chorus sounding alarms. A number of pieces from editorial boards and soundbites from politicians give the impression we’re a geopolitical move or two away from a Xi Jinping induced technological dark age.
“America today is dangerously reliant on foreign producers of semiconductors. America has receded while China has advanced. New chip-production capacity in the next decade would make China the largest semiconductor manufacturing location in the world…”The Wall Street Journal (July 26, 2021)
“America once led the world in making leading edge semiconductor chips. Today we produce 0% of those chips in America, 0%! That’s a national security risk and an economic security risk. The Department of Defense has been warning us for years…[We have to] protect ourselves. We are totally reliant on Taiwan and China for critical supply.”Gina Raimondo, Secretary of Commerce, to Congress on April 20th, 2021.
Furthermore, Republican Senators Tom Cotton and Bill Hagerty characterized the exporting of technology to Semiconductor International Manufacturing Corporation, China’s partially state-owned semiconductor fab, as a “clear national security threat,” based on close ties between SMIC and the Chinese Military. And yet, despite SMIC and Chinese telecom giant Huawei getting their hands slapped and earning a spot on the U.S. Government’s trade blacklist, an astonishing 301 export licenses with a value of $103 billion were greenlit for US firms by the Dept. of Commerce between April and October of 2021. Talk about the left hand not knowing what the right hand is doing.
Let’s also factor in the highly symbiotic relationship between the United States and China. They are our largest goods trading partner, with $559.2 billion in total (imports + exports) trade flowing between the two “enemies” as recently as 2020. They were our largest supplier of imported goods in 2020, and the US-China relationship supports 2.6 million jobs.
iPhones, Air Jordans, Ray Bans, Levis, and countless other quintessentially “American” things come from China. They are, as The Premier himself has said, “the world’s factory.”
And as economic conditions improve in the PRC, they’re expected to have a middle class that is larger than the entire US by 2026, and some experts predict China will be importing $520 billion or more of U.S. goods and services by 2030. Guess what they want? iPhones, Air Jordans, Ray Bans, and Levis. The economic opportunities that are available now, and the ones to come, are far too great for any American firm to ignore.
Adversaries, really? Maybe it’s better to think of this as two superpowers adjusting to a new technologically-driven, highly volatile world where each of them has strengths, weaknesses, and opportunities.
“Industry” Is Too Vague A Term
Before we can actually decide whether or not Chinese semiconductor manufacturing represents a threat, it would be helpful to take a look at the term “semiconductor industry” and understand why a more nuanced view of chipmaking paints a more accurate picture than what the soundbites describe.
Taking what is essentially sand, and a few other elements, and turning them into the brains of the electronics that power our world is a vastly complex, global, interconnected process. For our level of analysis, we can break it out into six different areas, with different business models, different degrees of technical complexity, value add, and resulting economic “power.”
Six Degrees of Semiconductors
Chip Design: Firms like Nvidia and Qualcomm that design but don’t fabricate chips. These “fabless” companies create 53% of value-add within the industry as they are responsible for almost all high-level innovation.
Integrated Circuit Fabrication: Foundries that are paid by fabless firms to transfer their chip designs onto silicon. Asia leads here, and it’s this imbalance that is the cause for so much alarm.
Chip Packaging: The highly-commoditized process of taking the imprinted-upon silicon and effectively “curing” it into a chip that’s ready to be packaged and shipped to an OEM.
Semiconductor Manufacturing Equipment: Tooling and machines used in all phases of production, but especially by foundries.
Electronic Design Automation: Companies who make software that helps chip designers automate the process of creating new IP.
Core IP: Businesses that make their money by selling premade blocks of software to the fabless companies to use in chip design.
Where We Stand In Each Sector
In short, the US and our allies dominate in all sectors of the semiconductor supply chain except one, and that’s foundries. In the categories that create the most value, innovation and technological differentiation between nations, we have a veritable monopoly.
Although the US is not as strong in the pure fabrication market as some would hope, we and our allies still control 73% of global fabrication capacity, and if we examine where that capacity is physically located, the answer is that 60% is outside of China and Taiwan. Factor in the nearly $1 Trillion in announced or planned semiconductor
Huawei’s tale has two sides–despite the licenses granted–the blacklist, which relies on our might in the design arena to work, hit the Chinese telecom giant hard, and denying them access to 5G chips and fabrication capacity in Taiwan caused years of declining revenues and forced Huawei to sell their Honor brand in an effort to stay afloat.
So that’s what happens when the US and our allies decide to throw a little weight around. It has an impact.
China Knows What It Needs To Do
Buffering his nation against impacts like that is one of the main reasons for the Made In China Act 2025 (MIC) exists. Let’s (finally) take a look at the broad strokes.
Announced in May of 2015, the MIC aims to upgrade China’s status in the global electronics value chain, specifically semiconductors, from its “world’s factory” status made possible by low labor costs and supply chain advantages, to a hub of self-sufficient high-tech innovation. The Chinese government wants to increase the Chinese-domestic content of core materials to 70% by 2025 with an estimated $300 billion or more earmarked to do so.
To achieve their goals, the government outlined five policies
- Lower tax rates for high tech firms.
- Incentivized mergers and acquisitions of and by foreign tech companies.
- Increased R&D funding by large manufacturing enterprises.
- Direct state funding of R&D.
- A roadmap with specific targets for outcomes like R&D spending share, environmental impact, productivity, and digitization.
Supporting the five policies are five initiatives:
- Constructing 40 R&D centers across China by 2025.
- Development of high end projects in 10 key industries.
- Developing world leading green manufacturing and sustainability.
- Adopting Smart Manufacturing particularly robotics and digitization
- Increasing domestic production of new materials.
The ten industries deemed most valuable to achieving the MIC goals are aerospace, biotech, information technology, smart manufacturing, maritime engineering, advanced rail, electric vehicles, electrical equipment, new materials, biomedicine, agricultural machinery and equipment, pharmaceuticals, and robotics manufacturing. All are traditionally areas where China lagged in the international market.
MIC doesn’t actually end in 2025, that’s just the first major goal post for the plan. By 2035 China aims to be a mid-tier world manufacturing power, and by 2049 a leading world manufacturing power.
Can They Do It?
Having an ambitious industrial plan and enough capital to make it happen is one thing, it’s altogether another to pull it off in the highly complex world of semiconductor design and fabrication.
It would be equally ambitious to presume it was within the scope and expertise of this article to critically evaluate the capabilities of high tech Chinese manufacturing. The criteria by which we judge the quality of a semiconductor changes rapidly as well. Moore’s law doesn’t really begin to cover it.
Nanometres (nm) is a word that gets used quite a bit in regards to a chip’s level of sophistication. A nanometer is one billionth of a meter, making each hair on your head about 90,000 nm wide, and those hairs grow at roughly one nm per second. For quite some time the length of a chip’s transistor gate, commonly measured in nanometers, defined it as cutting edge or legacy. Smaller transistor gates mean more processing power in a given space, so it makes sense that was our metric of choice.
But it’s not anymore. Nanometers have largely become marketing, with this piece from HP Wire doing a great job explaining why advances in technology mean there are other measurements, like density and pitch size.
That further complicates our evaluation of Chinese technical sophistication because all we’ve got to really judge it by is the now outdated standard of node size. And as the rest of the world rushes towards a 3nm chip, even though we know a lot of that is marketing, SMIC is set to maybe create a 12nm chip in 2022. That puts them years behind the market, and with China’s technological development being hindered by the blacklist, it’s very difficult to imagine them leapfrogging the rest of the world in the next decade.
Is It Working?
Seven years after the world first heard of Made in China 2025, have there been any tangible results? It’s difficult to say, given that Beijing controls the messaging around it closely.
Here’s what we do know: It’s been de-emphasized in communications for a while now,presumably in response to Western tensions and as the plan progresses into implementation. The strategy is evolving as well, with the government issuing hundreds and hundreds of documents addressing specific implementation techniques by 2018.
Then Trump won, and Covid-19 hit, with its epicenter in China, and it’s difficult to describe that economic impact. It’s hard to imagine the shutdowns and shortages that have defined the last two years didn’t derail MIC progress. The US embargo on Huawei did major damage, despite the huge number of allowances granted to US firms selling to it.
New fabs in America are years away, but the USA Chips Act is specifically designed to encourage domestic semiconductor fabrication, companies like Intel and Texas Instruments have plans for new plants, and TSMC just announced they’re going to double their production capacity. With this flurry of new construction, plus a growing desire for technological self-sufficiency, It’s easy to imagine a scenario where China’s strength in the pure foundry marketplace is rapidly diluted within a few years.
Things Will Probably Be Fine
That’s not to say all the governmental hand-wringing is for naught. SMIC has long-denied any connection to the Chinese Military, knowing that such a relationship would be disastrous for foreign investment. But China is a one party state that has ambitious plans to upgrade their technological capabilities. So, it’s understandable when those denials are met with skepticism.
China’s goals for themselves would be difficult, if not impossible, to achieve without at least some involvement from the international chipmaking community. Access to the best in equipment and design requires them to play nice with the US and its allies, and it’s not clear we have the stomach to shut them out of the tech export market, given the number of exceptions granted to the so-called blacklisting of Huawei.
Government investment in technology R&D doesn’t always equal results. Infrastructure is simpler–fill a pothole, fix a road, build a bridge. Those are binary tasks with easy to observe results. Sending money down the line and hoping it results in breakthroughs is another thing altogether.
We’ve gone down that road as a nation before, against the same supposed threat, with Sematech in the 1980s, and opinions on its efficacy are mixed at best. So, China’s ascendance to a leading spot in chipmaking is far from assured, even if they are riding a massive wave of money
A simpler, far safer approach that’s less likely to lead to unnecessary disruptions is that the US and EU should focus on their own lofty fabrication goals rather than China’s efforts to largely play catch up. Given how much fabrication capacity is already controlled by the two and their allies, if the USA and EU Chips Acts work, we’ll most likely continue to lead the world in the highest-value aspects of chipmaking.
And that’s a thought that should make it a little easier to sleep tonight.