China’s Quest for Semiconductor Self-Sufficiency

Executive Summary

This CETaS Briefing Paper examines how China’s quest for semiconductor self-sufficiency will affect the UK and Korean semiconductor industries. The paper builds on an April 2024 CETaS Research Report that assessed the ways the UK and Korea could bolster cooperation between their semiconductor and AI industries for mutual benefit.

Since October 2022, US-led export controls have explicitly targeted the Chinese chip industry, seeking to starve it of the critical components needed to accelerate progress in artificial intelligence (AI) and military innovation. These measures, while partly successful in slowing China down in the short term, have further emboldened Beijing in its quest for self-sufficiency and dominance in numerous parts of the semiconductor supply chain. Commentators have debated China’s strategic options for achieving its economic and political goals in a chip-deficient environment, but there is a need for greater focus on the second-order effects of China’s domestic ambitions on countries such as the UK and Korea. Despite not being at the forefront of ongoing US-China export control disputes, they still risk being caught in the crossfire of retaliation and deepening protectionism. 

China’s desire for semiconductor self-sufficiency predates recent geopolitical wrangling – the Made in China 2025 policy was instituted in 2015, paving the way for three iterations of The Big Fund to support self-sufficiency efforts. Crucially, the drive to achieve self-reliance is a whole-of-society undertaking in China, supported by vast amounts of centrally directed investment. 

However, as the semiconductor industry is highly R&D-intensive, an investment of hundreds of billions of dollars over several years is still no guarantee of short-term success. Some areas of the industry have benefited more than others, which will influence the type of leverage that China can exert over countries such as the UK and Korea. This paper focuses on five mechanisms through which China could gain leverage over them:

1. Intellectual property (IP) and chip design. ARM – a UK-based company – is one of the world’s top two core IP vendors. RISC-V is an open-source standard that some believe will rival ARM’s dominance in the long run. China has devoted significant resources to the development of the RISC-V ecosystem and sees this as an insurance policy should future US export controls extend to ARM products. If greater Chinese influence in the RISC-V ecosystem coincides with increased global adoption of the standard, this could create new dependencies on Chinese chip design – which could have important defence- and espionage-related implications further down the supply chain. Nonetheless, the nexus between chip design and national security goes beyond the role of RISC-V: countries such as the UK and Korea face more fundamental questions about trust in the security of chips designed by Chinese companies, even if they deploy ARM cores.

2. Compound semiconductors. These are a class of semiconductors particularly relevant for the AI, defence and electric-vehicle industries, accentuating the importance of sovereign capability. Compound semiconductors are a top priority for China, and its monopolisation of the supply of critical minerals such as gallium is a key lever of influence. It is crucial to continue supporting moves to increase UK domestic capacity in compound semiconductors, while leaning further into global compound semiconductor partnerships. In this, the UK’s initial agreements with countries such as Malaysia and Taiwan are positive recent developments. 

3. High-bandwidth memory (HBM) chips. The growth in demand for AI chips has driven greater Chinese interest in the advanced memory sector, and HBM is a crucial target for China’s latest iteration of The Big Fund. Although China’s leading memory-chip maker has begun producing second-generation HBM sooner than expected, China’s capability in many aspects of the HBM supply chain still lags behind Korea’s market leaders. As China continues to step up its ambitions in the area, it is conceivable that SK Hynix/Samsung plants in China will come under heightened pressure. This context, alongside the growth in demand for low-power HBM chips, presents an opportunity for ARM, which should collaborate with Korean HBM providers on the development of architecture and IP for these chips. 

4. Mature-node vulnerabilities. China’s strength in mature-node chips – those produced at 28 nanometres (nm) and above – is well-recognised, augmented by rapidly expanding markets for electric-vehicle batteries and solar cells. Members of the G7 (and other like-minded countries) have expressed concern that this could lead to oversupply in the global market and the dumping of certain chips that drives prices down – tempting UK and Korean companies to become reliant on them. This could leave Chinese chipmakers in control of a strategically important segment of the semiconductor market. 

5. Policy coordination. Recent years have seen even greater central oversight of all aspects of China’s semiconductor strategy. This has allowed for a more precise focus in areas such as semiconductor manufacturing equipment and has made it easier for Beijing to control domestic partnerships and collaborations. The designation of supply chain-wide responsibilities to companies such as Huawei makes it increasingly difficult for UK or Korean firms to know who they are collaborating with in China. The move has a wide range of national security implications, as China’s implementation of its military-civil fusion strategy continues apace. In this context, there are lessons for the UK and Korea in aligning science and tech policy coordination in the broader context of new security and defence agreements with other allies in Europe and Asia. 

The UK and Korea occupy very different positions in the global semiconductor supply chain, but both countries will be affected by China’s quest for self-sufficiency in their own ways. They should approach this challenge collectively rather than individually. In the context of deepening UK-Korea collaboration, and new relevant agreements with allies such as India and the EU, there is no better time for the UK to prioritise this partnership.
 

Policy Recommendations

Table 1. Measures to boost resilience against China’s semiconductor industrial policy

Introduction

Countries such as the US, China, India, Japan and Korea are investing tens of billions of dollars in their semiconductor industries to insulate domestic supply chains from external volatility. While recent financial incentives from the US and European governments have been encapsulated in flagship legislation – via the US CHIPS and Science Act and the European CHIPS Act – similar incentives in China sprawl across numerous vehicles, with some estimating that the Chinese chip industry has received more than $150 billion in government support since 2014.^[2] This period has been defined by near-limitless resources, a large and protected market, and a sophisticated global espionage strategy.^[3] In its quest for “national rejuvenation,” China is prepared to employ a whole-of-state approach, using a wide range of state levers to achieve its technological goals.^[4]

China’s resolve for indigenous innovation has only been strengthened following the intensification of US export controls. De-risking is also moving apace in China, with domestic Chinese telecoms providers ordered to replace chips from AMD and Intel in their infrastructure by 2027.^[5] China recognises that, to achieve its goals in domains such as AI, it must overcome a chip-deficient environment and quicken its drive for self-sufficiency across the semiconductor supply chain. 

By consolidating its semiconductor industry, China aims to secure technology leadership, find new growth sectors, create high-tech jobs and generate revenues that alleviate its debt issues. In the context of an impending macroeconomic slowdown, new drivers for prosperity are vital in maintaining the legitimacy and legacy of the Chinese Communist Party (CCP) leadership.^[6] There are also longer-term geostrategic drivers at play. If China becomes a semiconductor superpower, this would lessen its reliance on Taiwan in the sector and could affect Taiwan’s strategic value to the West. Separately, Chinese semiconductor self-sufficiency could augment the military ambitions of countries such as Russia, Iran and North Korea – as of 2023, China provided 90% of Russian microelectronics vital for advanced armaments.^[7] The possibility of China’s domestic chip industry acting as the guarantor for war economies such as Russia’s should raise significant concerns for policymakers in the West. 

More broadly, there is a need for a more rigorous study of the second-order effects of China’s quest for chip self-sufficiency on countries such as the UK and Korea. According to Boston Consulting Group and the US Semiconductor Industry Association, global 200mm+ commercial semiconductor fab capacity is projected to shift significantly by 2030. The US is expected to reach 14% of total capacity share, and China 21%. Korea, Taiwan and Japan are forecast to maintain relatively stable shares at 19%, 17% and 15% respectively. Europe’s share is predicted to stagnate at 8% over the coming decade.^[8] Across different chip technologies, these figures vary significantly. 

From a product perspective, China is expected to see substantial growth in logic chip production, particularly in the 10-22nm and >28nm nodes. Conversely, its capacity in memory chip production (NAND and DRAM) is projected to decrease significantly.

Figure 1. Global wafer fabrication capacity by technology category and region (2022–2032) 

Source: US Department of Commerce; SEMI; BCG Analysis.

The substantial projected increase in China’s capacity to produce mid- to low-tech logic chips over the next decade suggests that it is likely to establish a pathway towards a self-sufficient semiconductor ecosystem for many of its strategic industries, including autonomous vehicles and smart devices. This is particularly significant because most electronics and vehicles still rely on mature-node chips. 

Although China may not achieve full autonomy, its progress towards self-reliance in logic chip production is likely to have a significant impact on global semiconductor supply chains. It could fundamentally disrupt market dynamics: increased capacity in mature-node chips would raise the spectre of oversupply and significant price pressures, damaging the competitiveness of UK firms in the global market.^[9] Moreover, many of China’s domestic semiconductor industry priorities overlap with areas that the UK considers to be either existing strengths or avenues for future strategic advantage. This is also the case for Korea, albeit in different ways. 

The landscape is further complicated by impending changes in the White House. Under President Joe Biden, US export controls have had implications for UK companies: these restrictions affected exports of equipment from vendors such as Oxford Instruments, and high-end AI chips from developers such as Graphcore. The Biden administration maintained tariffs on Chinese semiconductors that were introduced under the first Trump administration.^[10] There is now building anticipation about president-elect Donald Trump’s approach to tariffs and subsidies in his second term. Further tariffs may lead to increased costs across all sectors that use semiconductors, which could unduly harm smaller economies such as those of the UK and Korea. In addition, if the Trump administration imposes restrictions on Chinese chips due to trust and security concerns, it could accelerate a bifurcation in global standards between China and the rest of the world.

This paper provides a detailed analysis of the Chinese semiconductor landscape and assesses its externalities for the UK and Korean semiconductor industries. The paper also suggests measures that the UK can implement to bolster its domestic strengths in the industry and to mitigate the potentially negative effects of China’s approach to semiconductors, both unilaterally and in cooperation with key partners such as Korea.
 

1. China’s Semiconductor Landscape

There is no doubt about the importance of science and technology (S&T) to the Chinese leadership’s bid to fulfil its global ambitions. In a speech at the Nationwide S&T Conference, Chinese President Xi Jinping emphasised this point:

“Although China’s S&T undertakings have made great progress, our original innovation capabilities are still relatively weak, some key and core technologies (关键核心技术) are controlled by others (受制于人) and there is a shortage of elite S&T talents. We must further enhance our sense of urgency, give further impetus to S&T innovation, and seize the commanding heights of S&T competition and future development.”^[11]

The reference to key technologies being controlled by foreign powers underlies a core tenet of Chinese policymaking in the 2020s: the centrality of self-sufficiency across technological supply chains. This section of the paper explores the implications of that push for self-sufficiency, examining the institutional vehicles the CCP has put in place to fulfil its aspirations. The paper then considers the areas of strength and weakness in China’s semiconductor sector.

1.1 Centrality of self-sufficiency: Made in China 2025

Semiconductors are crucial to China’s self-sufficiency goals because they sit at the intersection of national security and high-tech growth industries identified by the CCP.^[12] Indeed, as early as 2014, President Xi noted that “semiconductors [are] a core technology that China should produce domestically” and a key plank of China’s pledge to “mobilise all means … to wrest technological supremacy from the United States and other nations.”^[13]

That was the year China published its National Integrated Circuit (IC) Promotion Guidelines to accelerate support for its chip industry, laying the foundations for the publication of the Made in China 2025 plan a year later. The latter set the ambitious goal of 70% self-sufficiency in semiconductors by 2025 (a target China will fall short of by a significant margin) and provided the strategic framework under which the Big Fund would operate.^[14]

It is important to appreciate the scale of Chinese investment since 2014 and the influence the CCP exerts on the sector’s trajectory. Subsidies have flowed to the semiconductor industry from central, provincial and municipal Chinese governments, as well as state-owned enterprises in the tech sector. According to the Information Technology and Innovation Foundation, China “has probably invested the equivalent of the US CHIPS Act [in the semiconductor industry] virtually every year since 2014.”^[15]

Lower taxes, R&D tax credits, tariff reductions and exemptions for domestically produced or imported equipment are just some of the advantages provided to domestic firms. Moreover, subnational governments in China sell land to Chinese semiconductor companies at reduced prices.^[16] This is particularly relevant considering that 43% of registered capital across the Chinese semiconductor industry is either directly or indirectly owned or controlled by the Chinese state.^[17]

Therefore, it is no surprise that China’s 14^th Five-Year Plan (2021–2025 inclusive) upped the ante by designating semiconductors as an explicit strategic tech priority, requiring a whole-of-society effort to achieve self-reliance. For example, electric-vehicle makers have been asked to boost their purchases from local automotive chipmakers to hasten the move towards independence from Western imports.^[18]

1.2 The Big Fund III

The Big Fund (I, II and III) has acted as China’s most centralised instrument for directing money to the domestic semiconductor industry. The most recent iteration of the Big Fund dedicated an extra $47.5 billion to the sector, exceeding the $39 billion in direct incentives Washington introduced via the US CHIPS and Science Act.^[19] The table below sets out the key elements of the three iterations of the Big Fund.

Table 2. The Big Fund (2014–2024)

Table 3. Key industry players^[23] [24]

Table 4. Key stakeholders

Table 2 demonstrates the marked shift between Big Fund II and Big Fund III. The latest iteration focuses on ensuring that China has the resources to develop and build AI-capable semiconductors without outside help.^[25] This is likely to involve a dual approach combining the promotion of Chinese capabilities across the supply chain with simultaneous efforts to plug specific gaps in areas such as large foundries and HBM production.^[26] 

A broader strategic challenge that China will hope to meet with this investment is the reinvigoration of its national labs and innovation centres, by further integrating scientific research with commercial applications. Success in applied research relies on an awareness of changes in market dynamics. This synergy is central to China’s prospects of ensuring that the quest for self-sufficiency – an inherently costly endeavour – is underpinned by an economic rationale.

1.3 Areas of strength and weakness

The large investments described above will not necessarily translate into the development of key indigenous capabilities in the short term. To understand the type of market leverage China could exercise over countries such as the UK and Korea, one needs to assess current areas of strength and weakness in the Chinese semiconductor landscape.

Compound semiconductors 

Compound semiconductors made with Gallium Nitride (GaN) and Silicon Carbide (SiC) are vital not only for advanced industries such as next-generation mobile communications and electric vehicles but also for national security and defence purposes. This is due to characteristics such as their ability to operate with low power loss at high temperatures and voltages, which makes them suitable for high-current environments.^[27]

China has listed compound semiconductors as a development priority in its 14^th Five-Year Plan, partly because of their importance to the global electric-vehicle transition.^[28] China’s SiC sector experienced significant growth in 2023, earning the country greater attention from renowned integrated device manufacturers.^[29] Historically, China has accounted for 5% of the global SiC material market, but reports suggest that in 2024 more than 50% of global SiC wafer supply will come from China.^[30] Moreover, China provides 80% of the world’s supply of gallium, an element essential to compound semiconductors.^[31]

To bolster domestic innovation and competitiveness in this field, China’s Ministry of Science and Technology established the National Third-Generation^[32] Semiconductor Technology Innovation Centre. The aim of the body is to coordinate national resources, focusing on key technologies and applications, and fostering collaboration between various stakeholders through a “1+N+X” system:

• “1” Innovation Centre. An open platform with advanced infrastructure, talent and research capabilities.

• “N” Collaborative Centres. Established in key universities, research institutes, leading enterprises and national laboratories.

• “X” Innovation Units. Diverse cooperation models exploring breakthrough innovations. 

Complementing this effort, the Third-Generation Semiconductor Industry Technology Innovation Strategic Alliance (CASA) was established in 2015. This enhances innovation capacity by promoting collaboration between industry, academia and research institutes; upstream and downstream sectors; enterprises of various sizes; and central and regional entities.^[33] Both the Innovation Centre and CASA focus on technology development and commercialisation in key areas such as equipment, micro-LEDs, power-supply technology and materials. CASA actively promotes tests of platform construction and formulates national standards to drive industrial technological development. It has also established international cooperation initiatives such as the Asia-Europe Third-Generation Semiconductor Science and Technology Innovation Cooperation Centre, and is a participant in the CCP’s Belt and Road Initiative.

Design and IP

China has rapidly scaled up its chip design sector in recent years, with significant support for Chinese startups coming from the government and private investors (95% of investor transactions go to local startups).^[34] Many startups see it as more feasible to enter the chip design phase of the supply chain than the manufacturing stage, where entry costs can be in the billions of dollars. The Chinese authorities believe that greater chip design capability will “drag along” the development of downstream processes and upgrade the semiconductor system as a whole.^[35] This includes an effort to close the gap in AI chip design: with growing demand from China’s hyperscale cloud and consumer smart device market lowering barriers to entry, it is no longer surprising to see Chinese fabless firms preparing 7–5nm chip designs for AI or 5G communications.^[36] China’s large original equipment manufacturers increasingly engage in chip design to develop alternatives to server chips sold by US companies.^[37]

However, it is important to note that Chinese startups still need to collaborate with foundry services to implement prototypes and verify performance – specifically with TSMC, which has a significant lead over alternative foundries in this domain. It is uncomfortable for Chinese design startups to rely on TSMC given the geopolitical tensions over Taiwan. 

EDA has historically been a weak link in China’s supply chain but, for several decades, the country has targeted the area in the hope of accelerating the development of domestic alternatives to foreign EDA tools and core IP.^[38] For example, the Big Fund II invested $125 million in EDA and IP firms. And, in 2021, 12 Chinese EDA companies in mainland China raised more than $310 million in investment, a 54% increase from 2020.^[39] At a more strategic level, China launched the National EDA Innovation Centre in June 2023 with the goal of breaking the US stranglehold on EDA software. Yet, in the medium term, China will likely make significant progress largely outside the most advanced chip nodes.^[40]

To avoid dependency on technologies that might be subject to export restrictions, Chinese companies have been encouraging the adoption and promotion of open-source design technologies such as RISC-V.^[41] Currently, there are two main instruction set architectures (ISAs) for designing chips: x86 (IP owned by the US-based Intel) and ARM (a UK-based company with US IP). The following table presents a selection of notable Chinese initiatives in this space.

Table 5. Chinese RISC-V initiatives 

Nonetheless, despite high-level backing for RISC-V, Chinese companies continue to use ARM when possible, as it remains the de facto standard for many chips. But by investing in RISC-V now, China is preparing for a future in which Chinese companies cannot use ARM ISAs and cores.^[47]

Advanced packaging and chiplets

China has developed a strong presence in assembly, testing and packaging (ATP). As of 2021, China accounted for 27% of the world’s 484 ATP facilities. By August 2023, Chinese ATP firms held 38% of the market and included the five largest players.^[48] In the last two decades, much of the market share expansion of the leading Chinese outsourced semiconductor assembly and test (OSATs) vendors has come through the acquisition of foreign companies.^[49]

There is an increasing realisation within the semiconductor industry that packaging will be a key pillar of innovation. Packaging is crucial for the kind of high-power semiconductors needed for AI applications – a shortage of advanced packaging has been a critical bottleneck in the production of Nvidia’s AI chips.^[50] 

One innovation that is having an important impact on the pace of change in ATP is chiplet technology. In traditional semiconductor packaging, a single die is packaged, tested and integrated into a system. With chiplets, designers can break the system into smaller individual components that are then integrated using advanced packaging techniques. This is becoming increasingly popular in high-performance computing, data centres and consumer processors. 

In China, there has been an intense focus on supporting the domestic chiplet industry in the last year, partly due to a belief that chiplets could help China circumvent some export controls.^[51] For example, if China cannot buy or make a single piece of a powerful chip, it could connect less-advanced chiplets that it can produce to achieve similar levels of computing power.^[52]

However, there are many caveats to this. China’s strength and market share in traditional packaging and testing do not necessarily translate into dominance in advanced packaging or chiplets – this technology is more R&D intensive compared to conventional packaging. One key success factor is whether international alliances to standardise chiplet technology (with Chinese participation) will subsume domestic Chinese alliances aiming to do the same thing.^[53] This may influence whether chiplets will achieve the required level of customisability: to create synergies with chiplets, several types of chip need to have similar performance.^[54] Moreover, no chip designer has yet demonstrated a high-performing chip that integrates multiple chiplets fabricated on an older process node (e.g. 12/14nm).

It is still too early to tell whether chiplets can become a key driver of China’s semiconductor independence. Some industry observers contend that “while China has made some progress, it really doesn’t have that capability on the leading edge” of semiconductor ATP.^[55] The country is unlikely to gain a position of leadership in the next 5–10 years. This is due to its relatively closed-off market and industry structure, and to the efforts of the US to augment its own chiplet value chain. 

Talent 

Like most other countries in the semiconductor landscape, China faces a perennial challenge in acquiring and retaining talent. Following criticism of the central government’s Thousand Talents Plan, which aimed to attract foreign expertise through comprehensive incentives, various alternative programmes have emerged. 

One notable example is the Excellent Young Scientist Fund established by the National Natural Science Foundation of China. This programme prioritises individuals with experience in overseas universities, R&D institutions or corporate research centres, and provides substantial benefits. Similarly, the Ministry of Industry and Information Technology’s Qiming Plan and talent programmes from local governments are actively recruiting young talent in national strategic industries such as next-generation information technology, advanced equipment manufacturing, biotechnology and smart energy. 

The influx of foreign talent is crucial to China in two main ways. First, it significantly enhances manufacturing capacity via the transfer of tacit knowledge in cutting-edge fabrication processes. Second, these programmes play an integral role in cultivating China’s next generation of semiconductor specialists. This multi-level approach to talent acquisition and development underscores China’s commitment to building a robust, self-reliant semiconductor ecosystem, leveraging both international expertise and domestic innovation. 

2. China’s Market Leverage: The Impact on the UK and Korea

This section focuses on how China’s domestic policy will affect the UK and Korean semiconductor industries, as well as key Chinese players’ options to exercise market leverage. 

2.1 Intellectual property and chip design: RISC-V, ARM and national security

Table 5 details Chinese efforts to develop the global RISC-V ecosystem. Of the 10 billion RISC-V chips shipped in 2022, more than half were made in China.^[56] These investments are raising hopes in China that RISC-V could break the x86–ARM duopoly, with low cost, ease of customisation and energy efficiency making RISC-V attractive to chipmakers. Because the AI market is changing rapidly, companies need to develop various chips for different purposes, which can lead to an increase in IP costs. Therefore, ARM’s pricing policy could be an inadvertent driver of RISC-V’s growth. 

Previous CETaS research describes the strategic importance of ARM to the UK semiconductor landscape.^[57] The Emerging Technology Observatory’s Supply Chain Explorer shows the UK has 43% of global market share in core IP, largely due to ARM’s position as the top core IP vendor.^[58] Notable recent developments include ARM’s planned entrance into the AI chip market in 2025, with heavy investment from SoftBank and partnership with firms such as TSMC.^[59] This could boost the UK semiconductor industry and open up new markets for AI hardware. Ensuring that UK semiconductor companies have access to this technology could also boost the competitiveness of the domestic AI industry. 

Nonetheless, while the performance of RISC-V chips lags ARM in complex computing tasks and the RISC-V software ecosystem remains nascent,^[60] the gap is closing with increasing investment.^[61] China’s strengths in industries such as consumer electronics and electric-vehicle companies – in parallel with its increased use of RISC-V architecture – further reinforce the growing competitiveness of Chinese chip design.^[62] This could create new dependencies on Chinese chip design, particularly as the UK is importing record numbers of Chinese electric vehicles.^[63] It is also problematic given that Chinese chip design capabilities could have defence and espionage-related implications, and that organisations linked to the Chinese military are also developing and promoting RISC-V.^[64]

National security concerns have motivated some US lawmakers to pressure the White House into taking a more restrictive stance on US-China engagement on RISC-V.^[65] While some commentators view this approach as counterproductive,^[66] there are broader strategic implications for the UK. US-imposed restrictions on access or contributions to RISC-V could reinforce ARM’s position as the go-to licenser of processor IP.^[67] This could preserve the UK’s strategic advantage in the IP market, particularly if it leads to increased ARM licensing in China itself. However, the opposite could also be true: restrictions on RISC-V could merely accelerate China’s support for local semiconductor design capabilities and present longer-term challenges for ARM. Given that China’s strategies in this domain are predicated on the expectation that US export controls will eventually affect its access to US and UK IP, these are scenarios that UK policymakers need to prepare for now.

In any case, it is important to distinguish between the ISA and the implementation of the ISA. ARM is both a proprietary ISA and a form of implementation, whereas RISC-V is an ISA that only becomes useful once a company implements it as a processor core. This is particularly relevant in policy terms because the concern that a country could become dependent on Chinese chips based on RISC-V is largely a concern about cores. Consequently, it is unlikely that decisions made by the RISC-V standards body will introduce fundamental vulnerabilities that benefit China. 

Moreover, the nexus between chip design and national security goes beyond the question of RISC-V. It raises a more fundamental question about trust in the security of chips designed by Chinese companies. For instance, if a Chinese design firm shipped an advanced driver assistance system that incorporated ARM cores alongside a custom NPU and a security module, this would pose a similar level of national security risk as a scenario in which the firm replaced ARM processors with those based on RISC-V. 

China could also exercise market leverage through the theft of research or IP from foreign universities or companies. As the Intelligence and Security Committee of Parliament observes in its 2023 China Report, “by stealing IP, the Chinese save money on research and development, thereby lowering the overall ‘to-market’ cost so that they can undercut the original product and dominate the market.”^[68] This is also a consideration for Korean companies, albeit in more nuanced ways: Korean companies operating in China have received extensive support from the Chinese Government while contending with systematic semiconductor technology theft, forced technology transfers, antimonopoly investigations and various forms of industrial espionage.^[69]

2.2 Compound semiconductors

The AI and defence sectors are likely to see growing demand for GaN compound semiconductors. As China vies for leadership in GaN technology (see section 1 on China’s monopolisation of gallium supply), it may become increasingly challenging for the UK and Korea to secure access to GaN resources and develop domestic capabilities. 

In addition to new materials such as GaN, new technologies such as silicon photonics are expected to have an increasing impact on AI and quantum applications. Silicon photonics is the base technology required to build photonic ICs – devices that process light (photons) instead of electricity (electrons). Photonic integrated circuits (PICs) are already being developed for neuromorphic computing, mimicking the brain to produce energy-efficient AI functions. Combining photonic chiplets with conventional silicon chiplets will lead to a step change in functionality.

Figure 2 shows how these trends will involve diversification of chip technology, a key part of which will be the evolution of cloud-based AI systems to include distributed AI at the edge (where computation is distributed across multiple edge devices or nodes for large-scale processing).

Figure 2. Future trends in AI and quantum

Source: Authors’ analysis.

This is why recent moves to increase UK domestic capacity are necessary. They include the allocation of a share of £99 million in funding to Cardiff University to lead a new compound semiconductor manufacturing hub^[70] and a £22 million DSIT investment in two Innovation and Knowledge Centres (IKCs) in Southampton and Bristol. The Southampton IKC, known as CORNERSTONE, is the world’s only open-access silicon photonics foundry creating PICs. The Bristol IKC, known as REWIRE, focuses on wide-bandgap compound semiconductors to help the UK achieve its net-zero ambitions.^[71] In addition, the Ministry of Defence’s acquisition of Octric Semiconductors in Newton Aycliffe creates a sovereign UK capability in Gallium Arsenide,^[72] while a proposed expansion of the Newport Vishay chip factory promises to focus on R&D in compound semiconductors.^[73] 

Yet there is still work to be done. A handful of large deals account for a high proportion of total investment in UK semiconductor companies each year, and a relatively small number of companies receive consistent year-on-year investment.^[74]

This speaks to the importance of foreign investment and partnerships: China’s increased presence in the compound semiconductor space should drive greater collaboration between the UK and its partners. A recent memorandum of understanding between the UK’s Compound Semiconductor Catapult and the Sarawak Government in Malaysia is one example of this.^[75] Closer cooperation with the Taiwan semiconductor industry could be another counterbalancing measure. A report published as part of the UK-APAC Tech Growth Programme (which helps UK tech companies expand across the Asia-Pacific) highlighted the fact that Taiwan has a nascent midstream compound semiconductor industry but needs the world-leading technology being developed by many UK firms.^[76] Encouragingly, Oxford Instruments has already agreed on a cooperative research project with the Industrial Technology Research Institute in Taiwan for the development of next-generation compound semiconductors.^[77] 

The trajectory of the UK’s compound semiconductor industry will be heavily influenced by the extent to which it can balance the need for international collaboration and investment with the need to block certain deals on national security grounds. The former will help the country stay competitive vis-à-vis China, while the latter will deter investors.^[78]

2.3 High-bandwidth memory chips

From a Korean perspective, the most concerning aspect of China’s aggressive industrial policy is in the advanced memory sector. China has recognised the importance of HBM as a key component of AI semiconductors, particularly as it faces challenges in AI research and development due to US export controls on AI chips. This has led China to incorporate HBM into its plans for Big Fund III (see section 1). Indeed, China's leading memory chip maker, ChangXin Memory Technologies, started producing second-generation HBM this year, earlier than expected.^[79] While Korea’s SK Hynix currently leads the HBM sector, the gap with its competitors is narrowing as they pour more resources into HBM development. 

Nonetheless, China still has relatively immature capability in many aspects of HBM, such as DRAM core-die quality, die assembly and advanced packaging. Although China recently began mass production of HBM2, there will be a large gap between the performance of its memory devices and those of more advanced providers in Korea and the US. 

From a security perspective, it is notable that SK Hynix maintains key factories in the Chinese cities of Wuxi and Dalian. The Chinese Government could ostensibly target these facilities with measures such as hostile takeovers of Korean SMEs that have engineering technology essential for HBM and aggressive talent recruitment. 

Korea needs to establish a comprehensive technology protection framework for HBM. To maintain and expand its lead in the sector, Korea’s Government and companies need to match or exceed China’s large-scale investments in next-generation technologies^[80] (SK Hynix has already announced significant HBM investments through to 2028).^[81] For a dual approach focused on protection and promotion, it will be vital for Korea to cultivate a trusted and mature innovation ecosystem with reliable partners such as the UK. 

In this regard, it is notable that the boom in on-device AI is leading to growing demand for low-power HBMs from domestic companies such as SK Hynix and foreign firms such as Qualcomm and Meta. This presents an opportunity for ARM, which will need to collaborate with top-tier HBM providers to accelerate the development of architecture and IP related to low-power HBMs, providing a clear avenue for UK-Korea partnership in the sector.

2.4 Mature-node vulnerability and its national security implications

The mature-node chip market is far more pervasive across global supply chains than technologies such as RISC-V and compound semiconductors, increasing the importance of China’s leverage in the domain. China’s share of global mature-node production is expected to grow from 31% in 2023 to 39% in 2027.^[82] Rapidly expanding markets for electric vehicles, batteries and solar cells will further augment this mature-node strength, providing the impetus for China to establish many new mature-node chip factories.^[83] TrendForce reports that China is planning or constructing at least 32 large-scale wafer fabrication plants, which focus on the 28nm+ processes widely used for commodity chips in household appliances and the car industry.^[84]

Countries such as the US, the UK and Korea are worried that China will use the resulting oversupply to dump certain types of chips on the international market, driving prices down and inducing foreign manufacturing firms to buy more of them.^[85] Domestic subsidies allow Chinese semiconductor firms to compete in markets without the need for market-based rates of return, severely disrupting the established economics of innovation in the industry. Normally, firms depend on profits from one generation of semiconductor products to finance R&D for the next generation, but aggressive subsidisation by the Chinese state upends this.^[86]

The European Commission has responded to concerns about this by beginning discussions with European chipmakers on China’s growing production of commodity silicon.^[87] If Chinese production prices Western manufacturers out of the market, China’s chipmakers could gain control of a strategically important segment of the global semiconductor industry. This issue is strategically important partly because mature-node chips are constantly being refined in line with new requirements and applications – and are foundational to the aerospace and defence sectors.^[88]

In the medium-to-long term, China’s mature-node ecosystem will likely gain a strong foundation relative to the one TSMC currently has in leading-edge nodes today. Given that the mature-node market tends to be less profitable than that at the leading edge, the Chinese Government’s subsidy policy is a crucial asset in this endeavour. 

2.5 Policy coordination

This paper covers the various plans, funds and advisory groups that China has established in support of its semiconductor industry. By early 2023, however, the Chinese Government determined that these investments and the “scientist-led approach” were not producing the desired results.^[89] It called for even greater top-down policy coordination, leading to the establishment of the new Central Science and Technology Commission under the direction of Vice Premier Ding Xuexiang. Overseeing all aspects of China’s semiconductor strategy, the commission is tasked with improving the coordination of R&D between the state and industry, and with directing further investment in basic research and talent development.^[90] 

The accelerated effort to gain independence in the semiconductor manufacturing equipment space is an example of this top-down approach in action. In the first half of 2024, China spent more on chipmaking equipment than the US, Taiwan and Korea combined.^[91] The Chinese leadership recognises that it is particularly vulnerable in this domain, due to the monopolies held by Western firms such as ASML. The promotion of partnerships and collaboration between Chinese manufacturing equipment firms and fabs is one way to spur indigenisation efforts, but Chinese manufacturing equipment companies are struggling to find buyers that are open to collaboration on the refinement process.^[92] To avoid such bottlenecks, centralised mechanisms can provide incentives to ensure these companies prioritise domestic rather than international collaborations. 

Another example of policy coordination in action is the Xinchuang National Project, which aims to localise servers and databases. China’s Government and state-owned enterprises will purchase the products of this project, while all related policies will be led by the National Development and Reform Commission. 

A greater level of policy coordination allows the Chinese Government to allocate strategically significant roles to individual companies. Huawei is a prime example of this: its vertical integration of the supply chain and close ties to many companies within that chain make it a key coordinator of China’s lithography supply chain.^[93] Meanwhile, Huawei’s development of Kirin chips – which are suitable for use in Chinese-made electric vehicles – help address shortcomings in China’s capacity to develop automotive-grade chips.^[94] And the company’s work on developing AI chips (the latest being the Ascend 910C) poses an increasing challenge to the status of Nvidia’s leading products in China.^[95]

Striking the right balance in Korean industry’s relationship with China remains a significant challenge: factories in China account for 40% of total capacity for producing Samsung’s flash memory chips and 40–50% of SK Hynix’s DRAM chips.^[96] SK Hynix will also be wary of an anticipated strengthening of US export controls on HBM:^[97] its fabs are already unable to import extreme ultraviolet equipment, which will gradually push them towards producing older chips.^[98]

Additionally, for companies in the UK and Korea, decisions on cooperation and supply become more challenging if they cannot verify their business partners. This is especially germane given the well-documented ties between Huawei and China’s military and security apparatus.^[99] As the CCP tasks more companies with the achievement of strategic goals, and as China accelerates the implementation of its military-civil fusion strategy, it will become increasingly difficult to identify which parts of China’s supply chain are safe to engage with.^[100] This could have significant long-term implications for businesses in the UK and Korea that currently rely on collaboration with Chinese partners: 5% of the UK’s major semiconductor firms have parent companies based in China.^[101]

Figure 3. Parent company location of 61 major UK semiconductor firms

Sources: Mapping (De)Globalization; Fame132; Orbis Database133; Refinitiv Eikon.
 

3. Conclusion: Boosting Resilience and Reinforcing Strengths

The UK government’s new industrial policy plans will need stability and certainty for innovation, delivered through a long-term approach to supporting public and private R&D, bolstering skills and growing exports.^[102] This requires progress on long-standing innovation weaknesses, such as an inability to recruit enough diverse talent; a lack of access to finance for capital expenditure; and an overreliance on overseas manufacturing and prototyping facilities, which risks IP leakage.^[103]

The challenge is accentuated by the scale of China’s investment in its technology and manufacturing industries. The country has continually exceeded the expectations of market-watchers, with the result that Korean factories in China are facing pressure from local competition quicker than they expected. In recent years, Korea has anticipated this challenge by designating various national critical technologies and reinforcing protections for SMEs and innovative startups through incentives. 

The US CHIPS and Science Act R&D programme popularised the “Lab-to-Fab” motto – bridging basic technologies to mass production and the market. The UK has strengths in the “Lab” (i.e. fundamental research) while China has strengths in the “Fab” (i.e. manufacturing). For the UK to address its shortcomings on the latter side of the equation and expand its current lab-based capabilities into the proof of concept/prototype area, it will need to deepen its partnerships with countries such as Korea, Taiwan and Japan. 

Optimisation is a key success factor in the semiconductor industry, affecting every link in the value chain – from design to development, mass production, packaging, and testing and verification. Korean semiconductor companies have an advantage in optimisation because they operate a business model based on integrated device manufacturing. In the era of AI competition, the semiconductor value chain is becoming fragmented once again, thereby weakening optimisation. If the UK and Korea are to collaborate successfully in the face of these challenges, they will need to keep optimisation front of mind – bridging the gap between the UK’s strengths in design and IP and Korea’s strengths in manufacturing and memory devices.

References