Manufacturing Update - 2 October 2024
Insights from articles and reports of interest on manufacturing technology, management, policy, and economics in the US and abroad
“Manufacturing Update,” is a periodic newsletter excerpting and summarizing a small collection of recent articles on manufacturing technology, management, economics, and policy developments in the US and abroad that we think will be of interest.
The concept is to update the group on manufacturing developments that may be of interest, to share ideas, and to promote fluency and expertise with manufacturing issues and the debates around them. We encourage readers to send articles that you think will be of interest to us at mfg-at-mit@mit.edu.
From broad unilateralism to narrow multilateralism in semiconductor chip controls
Intel’s strategic transition: can new orders and cost cuts lead to a rebound?
Can Intel Catch TSMC?
Chinese Solar and the German Paradox
China Is Rapidly Becoming a Leading Innovator in Advanced Industries
Geopolitics and the changing landscape of global value chains and competition in the global semiconductor industry: Rivalry and catch-up in chip manufacturing in East Asia
1. “From broad unilateralism to narrow multilateralism in semiconductor chip controls,” Douglas Fuller, Danish Institute for International Studies (DIIS) Policy Brief, September 12, 2024
Almost two years into Biden’s October 2022 broad and unilateral chip controls on national security grounds, there is evidence of significant success in limiting China’s expansion of advanced fabrication (i.e. chipmaking) for the next decade.
Success has come from refining policy and coordinating with key allies, especially Japan and the Netherlands. There are still concerns about China’s expansion of legacy chips and access to AI chips.
The Huawei story as a case study: In August 2023, many hailed Huawei’s Mate 60 as proof that China had broken America’s technology blockade on Chinese advanced chipmaking efforts. Huawei’s procurement of 7nm logic chips from China’s Semiconductor Manufacturing International Corporation (SMIC) – when controls were supposed to prevent China from producing even less advanced 16/14nm logic chips – supposedly demonstrated China’s independence from American chipmaking equipment. However, everything was not as it appeared. SMIC’s fabrication factory – ‘fab’ for short – was full of American equipment and had new, licensed supplies for such equipment – occasionally by legally dubious means – despite the controls. America’s unilateral controls seemed to be a failure.
In reality, by January 2024 the US introduced effective multilateral controls. While the most advanced extreme ultraviolet (EUV) lithography – the process by which chip circuitry is imprinted on wafers by lasers – had already been controlled by the multilateral Wassenaar Arrangement, SMIC had used the most advanced of the previous generation of deep ultraviolet (DUV) lithography equipment, supposedly the target of American controls, to make Huawei’s chips. The downside to using this older DUV equipment was that the yields were much lower. China would have to expand beyond SMIC’s existing capacity to produce the large quantities of chips needed for Huawei’s smartphones and AI processors.
Looking at Huawei’s smartphone sales and AI chip design decisions reveals how tight advanced chipmaking supply remains in China. Gavekal Research estimates that SMIC, with its low yields, can only produce 8.75 million 7nm chips per month. This estimate is consistent with Huawei’s low sales figures for the Mate 60 and subsequent advanced smartphones, despite large demand for such phones in China. Since the launch of the Mate 60, Huawei has on average produced only 3.7 million units per month – a quarter of its pre-controls sales of advanced smartphones – and 3.9 million units monthly in 2024 Q1, which would represent almost half of SMIC’s capacity.
Toward Multilateral controls. The imposition of hard constraints on China’s advanced chipmaking capacity is a surprising success compared to the very loose licensing of most of 2023 including equipment inputs that made it to SMIC’s advanced fab, which was already on the US foreign “Entity List” as well as affected by its broader licensing regime. The move from a loose to tight export control regime was made possible by working with allies. The US was able to use its leverage as a chip market to get agreements from companies in the Netherlands and Japan.
With the Netherlands’ ASML providing the bulk of lithography equipment in the world, and Japan providing most of the remainder as well as a wide range of other equipment, it was critical for the US to get these two allies on board. ASML was in a vulnerable position because its lithography equipment had significant American content and thus placed it under the long arm of American export controls. Nonetheless, the Dutch–American relationship was not just a one-sided imposition of controls. While the Dutch agreed to implement controls and wrote them in 2023, setting January 2024 as the start date of these controls, Americans learned from and built on the new Dutch technical regulations in revising their own lithography controls.
US had little leverage over Japanese fabrication equipment firms as they were not as reliant on American technology inputs as ASML. Americans also confronted Japan’s very loose export controls based more on requiring reporting exports than on seeking licenses. However, being located in East Asia and having some territorial disputes with China, Japan shared America’s concerns security concerns. Japan moved to implement a strict licensing regime but made the US compromise too. The Japanese rules still allow China to buy equipment that could be used for the second-most advanced generation of logic chips targeted by export controls, FinFET (used in 14-10nm logic chips), but not the newest generation of gate-all-around logic.
Overall, US multilateral chip controls are working to date.
Excerpted with edits; more at: https://www.diis.dk/en/research/from-broad-unilateralism-to-narrow-multilateralism-in-semiconductor-chip-controls
2. “Intel’s strategic transition: can new orders and cost cuts lead to a rebound?” Judy Lin, DigiTimesasia, September 18, 2024
Many still regard Intel as one of the greatest American companies. However, recent struggles have led to a less flattering label: "too big to fail." The critical question now is whether Intel can see a glimmer of hope amid news of new orders and ongoing efforts to cut costs by $10 billion, including a significant reduction of 15,000 employees from its global workforce and real-estate development exits. Intel's recent announcement of a strategic alliance with Amazon Web Services (AWS), along with $3 billion in CHIPS Act funding from the Defense Department, underscores the robust support Intel receives from both the government and the private sector. With a historic dominance in the CPU market—maintaining a 76% global market share as of the first quarter of 2024—it is challenging to envision such a giant stumbling. Yet, persistent financial challenges have plagued its operations and led to stagnant sales growth.
In a recent 10-Q report filed on August 2, Intel revealed that its Client Computing Group experienced a 19% revenue growth in the first half of 2024, while areas like Data Center and AI, Network, and Edge showed little change. Although Intel Foundry's sales rose slightly in the second quarter, overall sales for the first half of the year decreased by 3%. Intel Foundry consists substantially of process engineering, manufacturing, and foundry services groups that provide manufacturing, test, and assembly services to Intel Products business and third-party customers.
Taking a closer look, recent figures indicate a concerning trend: external customers are moving away from Intel Foundry Services (IFS). Sales contributions from these external parties dropped by more than 30% in the second quarter and the first half of this year compared to the same period last year while losses widened. Losses from its foundry services subsidiary totaled $2.8 billion in the second quarter of 2024. Intel faces stiff competition from Qualcomm, a newcomer in the AI PC market, and the pressure is intensifying in the X86 architecture landscape. However, the challenge of IFS is greater than anything else.
CEO Pat Gelsinger has made extra efforts to maintain employee confidence in IFS, emphasizing the importance of building momentum as the company approaches the launch of Intel 18A, its pending new chip. He highlighted the need for greater capital efficiency within this division while also focusing on creating a more competitive cost structure to achieve a $10 billion savings target.
Exerpted with edits; more at: https://www.digitimes.com/news/a20240918VL205.html
3. “Can Intel Catch TSMC?” David Manners, Electronics Weekly, September 11, 2024
When you get behind it takes 10 years to catch up,“ said Malcolm Penn CEO of Future Horizons, “you’re constantly chasing a moving target.” Intel is trying to do it in four years. “Technically they should be able to do it if they could get the right people,” added Penn “but engineers like to work for successful companies They do not want to worry about whether they’ll still have a job in a year or two.”
TSMC has been running its 3nm process for a year while Intel’s best process in volume production is Intel 4 at Leixlip. Intel 3 and TSMC N3E are the final finfet generations for both companies and that is when a chance opens for Intel to catch up as the first Gate-All-Around (GAA) processes from both companies go into production. Intel has a more advanced GAA process than TSMC’s GAA process and Intel’s is planned for next year, but it’s unclear whether it will be ready for mass production next year. According to Broadcom, it is not ready yet.
Penn pointed out Intel took control of the company from engineers and that was when its process capability declined. He reckoned that Intel is now about a year behind TSMC in process capability.
Excerpted with edits; more at: https://www.electronicsweekly.com/blogs/mannerisms/manuf/can-intel-catch-tsmc-2024-09/
4. “Chinese Solar and the German Paradox,” Kyle Chan, High Capacity, April 20, 2024.
How is it that countries with pioneering technology, early market adoption, sophisticated industrial ecosystems, and abundant capital get overtaken by China in so many industries? In solar manufacturing, Germany at one time seemed to enjoy unbeatable advantages only to later have Chinese firms run circles around them. While Germany seemed poised to dominate solar, it was Chinese solar firms that used German equipment to make solar cells and modules to sell back to Germany's subsidized market. What happened?
In the 2000s, China’s solar industry was completely transformed by two game-changing developments in Germany. (Other countries were involved too, but Germany’s role looms large.) One was a change in Germany’s renewable energy policy—the Renewable Energy Sources Act—that triggered a global surge in demand for solar power. Specifically, Germany established feed-in tariffs, which guaranteed substantial revenues for electricity from renewable sources, such as solar. This caused a boom in demand for solar modules and cells. At its peak, Germany made up half of the world’s new solar installations from 2004 to 2010.
The other big development around the same time was the rise of German “turnkey” solar manufacturing equipment. German firms like Centrotherm started selling sets of PV manufacturing equipment as complete all-in-one production lines. German technicians would even set up all the equipment and train a company’s staff how to use them. This enabled firms without deep expertise in solar technology to quickly jump into PV module and cell manufacturing.
German firms were never able to scale up as quickly and reduce costs as aggressively as their Chinese competitors. Despite their obvious advantages, German solar manufacturers were beaten at their own game by Chinese upstarts.
How did China do it? Aside from feed-in tariffs and turnkey manufacturing equipment, let’s look at some possible explanations:
China’s scientific knowledge base in solar. China’s efforts to develop solar photovoltaic (PV) technology go back to the 1950s and 1960s when Chinese researchers were studying solar power for space satellites (like the US did). China’s R&D efforts on semiconductors also helped position it later to produce polysilicon and make solar manufacturing equipment. Yet during the early boom years, the country was short on talent, and Chinese firms had to recruit Chinese PhDs trained at foreign universities, paying them top dollar. Germany was in a far better position in terms of both scientific research and industry expertise.
Labor costs. Lower wages helped Chinese solar module manufacturers cheaply turn imported German solar cells into modules, which they sold back to Germany and other markets. But solar cell production has long been highly automated, including in China. The real cost constraint for solar cells was equipment and inputs. Germany had an advantage in both these areas as a producer of manufacturing equipment and a producer of polysilicon, the primary material in solar cells.
Financing and subsidies. China’s central government provided billions of dollars of credit to support the solar industry, particularly through the China Development Bank. But this didn’t start until 2009 after Chinese solar firms already dominated the global market. Instead, Chinese firms relied on two other sources of financing. One was local government support at the provincial or municipal level. For example, the city of Wuxi gave Suntech’s founder financing and support to move from Australia (he later thanked Wuxi’s party secretary for his success). While US and German solar firms accused Chinese firms of unfairly benefiting from subsidies, the US and European countries also provided their own solar industries with subsidies and other support.
US capital markets also provided financing. Chinese solar firms like Suntech, Trina, Yingli, Jinko, and JA Solar raised hundreds of millions of dollars listing on the NYSE or Nasdaq. (Many of these stocks later tanked or were delisted when the solar market crashed.)
Solar demand in China. Today, China is by far the single largest market for solar energy, adding more solar capacity in 2023 alone than the US ever added. But interestingly, China’s surge in solar manufacturing predated its surge in domestic solar power capacity, which took off with China’s own feed-in tariff program in 2011. So while China’s own renewable energy policies would later help its solar manufacturers survive during a difficult period for the global solar industry, it was demand outside of China—namely, in Germany and other parts of Europe—that launched China’s solar industry.
Speed and nimbleness. The speed and scrappiness of Chinese solar firms is also a cause (as explored in recent work by researchers Greg Nemet and Kelly Gallagher).
Supply chain dominance. China currently dominates every part of the solar supply chain.
Today we’re witnessing various iterations of the German paradox unfold in real time. Chinese EVs and batteries have gone from being mocked for their poor quality and crude design to being feared as a global threat.
Excerpted with edits; more at:
5. “China Is Rapidly Becoming a Leading Innovator in Advanced Industries,” Rob Atkinson, ITIF–Hamilton Center, Sept. 16, 2024
There may be no more important question for the West’s competitive position in advanced industries than whether China is becoming a rival innovator. While evidence suggests it hasn’t yet taken the overall lead, it has pulled ahead in certain areas, and in many others Chinese firms will likely equal or surpass Western firms within a decade or so.
China has reached a new stage in its economic development, with much greater innovation capabilities in its universities and domestic companies. On many innovation indicators, China now leads the United States.
China leads or is on par with global leaders in commercial nuclear power and electric vehicles and batteries. It lags behind in other key sectors, including robotics, biopharmaceuticals, chemicals, and AI. But it is making rapid progress.
The combination of low costs and growing innovation capability make an increasing number of Chinese companies formidable global competitors.
This rapid innovation progress stems from the Chinese Communist Party’s determined effort to dominate global markets in a host of advanced industries.
While the Chinese innovation system is not perfect, it is much stronger than previously understood—and there are many aspects of it the United States should emulate.
To compete with these policies, the US government must identify key sectors that are critical for national and economic security and invest adequately to win. To do so, America should create industrial research institutes, a competitiveness-oriented DARPA, and an industrial development bank; triple the research and experimentation tax credit; and institute a seven-year, 25 percent credit for capital equipment.
Excerpted with edits; more at: https://itif.org/publications/2024/09/16/china-is-rapidly-becoming-a-leading-innovator-in-advanced-industries/
6. “Geopolitics and the changing landscape of global value chains and competition in the global semiconductor industry: Rivalry and catch-up in chip manufacturing in East Asia,” Chan-Yuan Wong, Henry Wai-chung Yeung, Shaopeng Huang, Jaeyong Song, and Keun Lee, Technological Forecasting and Social Change, Sept. 2024
There is a changing landscape of global value chains (GVCs) and competition in the global semiconductor industry in the context of new geopolitics. This has featured the United States implementing “chokepoint” measures to limit the rise of semiconductor manufacturing in China. Overall, the paper finds that these US measures, like the IRA and CHIPS act, will have important impacts on semiconductor GVCs, especially in three types of memory (DRAM and NAND and logic chips), and will slow down the speed of China's ascent.
China's is quickly catching up in producing large quantities of chips, but slow in improving quality except in High Bandwith Memory (HBM) chips, which is the most recent segment where China has already surpassed Korea and Taiwan in terms of the number of patents. Whereas China has been catching up rapidly in the number of patents, it might encounter problems in turning that into larger market share given the existing restrictions in accessing complementary technologies and chipmaking equipment, such as advanced lithography machines (EUV) or even more matured technologies (DUV), and software.
Severely constrained by these barriers, China progress is faster in lower-end products by foundry firms (e.g. its company SMIC), medium to high in NAND memory chips (e.g. YTMC), and slow or difficult in DRAM (e.g. CXMT). In the meantime, China has been making progress in domesticating value chains in diverse equipment and components in chip manufacturing.
Excerpted with edits; more at (paywall): www.elsevier.com/locate/techfore
Since 2022, MIT has formed a vision for Manufacturing@MIT—a new, campus-wide manufacturing initiative directed by Professors Suzanne Berger and A. John Hart that convenes industry, government, and non-profit stakeholders with the MIT community to accelerate the transformation of manufacturing for innovation, growth, equity, and sustainability. Manufacturing@MIT is organized around four Grand Challenges:
1. Scaling advanced manufacturing technologies
2. Training the manufacturing workforce
3. Establishing resilient supply chains
4. Enabling environmental sustainability and circularity
MIT’s Bill Bonvillian and David Adler edit this Update.