Semiconductor manufacturing: What China really wants

This summer, Her Majesty’s (HM) Government launched a national security review of the People’s Republic of China’s (PRC) purchase of the Newport Wafer Fab (NWF), a semiconductor factory in South Wales. A final decision on the acquisition by state-backed electronics company Wingtech through its Dutch subsidiary Nexperia is expected in January 2022. 

At almost the same time as HM Government announced that Sir Stephen Lovegrove, the National Security Adviser, would review the purchase, a report[↗] from the White House called the circuits that firms like NWF manufacture the ‘DNA’ of the high-tech economy, which underpin ‘state-of-the-art military systems’.

Gina Raimondo, President Biden’s Commerce Secretary, has asked allies to help slow the PRC’s innovation rate in the sector. Chinese strategy is quite simple[↗]: rapidly increasing its control of semiconductor supply chains. This has deep national security implications. 

Yet, claims that democratic countries’ positions in the industry have been ‘stolen’ by East Asian competitors are often based on a misunderstanding of a vast sector that includes many different activities. The industry is extremely complex, but firms broadly fit into particular niches[↗]: first, the big intellectual property-holders are the design firms; next, foundries (or ‘fabs’) produce integrated circuits (ICs), usually manufacturing for the design firms; packaging and assembly companies generally package these circuits in simpler cases, as components for products like smartphones. 

Separately, manufacturing-equipment companies make the kit that all these firms use for manufacture and design. The trend is strongly towards specialisation, with a few integrated device manufacturers (IDMs) like Intel and Samsung that design, make and package. But the biggest value-add is in the design firms: packaging trails a distant third[↗].

So where is the PRC? This is surprisingly straightforward. Design is now dominated by one German and three American firms; Taiwan dominates fabs, followed by South Korea (Taiwan Semiconductor Manufacturing Company (TSMC) is now well ahead of any competitor technologically, and with Samsung has 81% of the world market, hence the claims that East Asia has devoured the semiconductor industry); and the bulk of manufacturing equipment comes from the United States (US), the Netherlands, and Japan. The PRC’s position in all these fields is negligible. 

The ‘assembly, testing, and packaging’ segment is Taiwanese-dominated, but now also features some major Chinese players[↗] such as Tianshui Huatian, JCET Group, and Tongfu Microelectronics. This sector, less capital- and skill-intensive than the others, was the first to become dominated by East Asian firms. The PRC now hopes that the sector will provide a foot in the door.

Yet, the American market share of around 50% of the industry as a whole is ten times the PRC’s. The PRC has almost nothing in the highly profitable fab-less design field and lacks Taiwan’s ability to make leading-edge 3-nanometre (nm) chips. 92% of manufacturing capacity for 10nm or under is Taiwanese[↗]. So what’s the problem?

Some analysts[↗] say the PRC is now gaining ground and may be just six years behind the leaders in 7nm and 12nm technologies. Meanwhile, the concentration of manufacturing in Taiwan, where TSMC alone has 55% of the world market, is now a serious geostrategic risk. The PRC’s market shares are low but growing fast, and its purpose is not to make a profit. The White House warned allies this year that ‘armed with billions in subsidies, the Chinese government [is] on a buying spree for foreign semiconductor companies’. 

One example of Chinese state investment is the China Integrated Circuit Industry Investment Fund (CICIIF), aka the ‘Big Fund’ (大基金), launched in 2014. The Big Fund[↗] raised US$51bn in 2015 alone, largely from state-owned sources. The PRC’s self-reliance plans cover ‘the whole IC supply chain[↗]’.

The Chinese state helps its companies stay in business although ‘most do not appear to be making a profit’, while they are securing ‘access to and control of’ semiconductor intellectual property. The White House does not mention that this increasingly comes from the UK.

The fundamental military technology

Back in South Wales, NWF’s products are not cutting-edge, but they are in demand, especially for their promising use in electric vehicles, where the PRC is the largest market. 

But NWF reportedly has contracts[↗] to develop chip technology for other purposes, including with Cardiff University on technology for fighter aircraft radar, a project intended to supply partners such as missile developer MBDA. Chinese military-linked organisations have an interest[↗] in UK radar research and next-generation jet engines. These are priorities for the PRC as it improves navy and flight capacities. This does not imply that Wingtech intends to put NWF’s technology to military use, but Beijing’s ‘military-civil fusion’ strategy means the Chinese government can demand that civilian firms share technologies with the military regardless of any company’s wishes. (Missile guidance was the first military field[↗] to benefit from microelectronics: the Minuteman II ICBM was the first weapons system to employ ICs.)

Yet, Wingtech is not simply a private firm, but a state-private hybrid. Analysts at Datenna have traced[↗] 30% of its shares to PRC government entities. The owners of the UK’s largest wafer fab now include four Chinese cities and one province. Wingtech’s largest private shareholder is ‘Lhasa Economic and Technological Development Zone Wentianxia World Investment Co Ltd’, owned by Zhang Xuezheng[↗], the Chief Executive Officer of Wingtech and Nexperia. The Hong Kong Trade Development Council reports[↗] that Tibet Zhongkai Mining Group was an early investor in this company: according to[↗] Radio Free Asia, security forces have confronted Tibetans protesting against its digging plans at sacred Tibetan sites.

The PRC’s strategy is wasteful and generates little intellectual property, but is designed for maximum leverage. HM Government has been told that the PRC’s purchases of firms like NWF are not security concerns because these firms are not at the cutting-edge. But this fails to grasp how the PRC’s drive for techno-economic power is informed by much broader considerations. 

The NWF purchase exists alongside a coordinated programme involving different parts of the PRC’s state and military. Its first purpose is to expand control of the supply chain; the second aim is rapid military expansion. In 2018, a Chinese firm backed by a military-civilian fund bought Finnish company Beneq; Dutch firm Ampleon began focusing on defence and aerospace after it was purchased in 2015.

In the 20th century, access to oil was a prerequisite for military power; the PRC’s intentions for every firm will never be fully known, but this century military innovation will depend on advanced semiconductor technology. NWF is a non-exceptional fab, but the Chinese military also needs volume. Weapons systems ‘do not rely on the most advanced foundries’: ICs etched at 28nm to 350nm ‘find many industrial applications in the automotive, robotics, and armament sectors[↗]’. 

Chinese companies like SMIC deny producing integrated circuits for the Chinese military: but the US has told its firms that transferring technology to firms like SMIC poses an ‘unacceptable risk[↗]’ of diversion to ‘military end-use’ (since 2020, the US has deemed SMIC a ‘Chinese communist military company’, updating this list to cover growing swathes of China’s semiconductor industry and its end-users).

Regardless of NWF’s capacities, Australia’s experience shows the risks of crucial supply chains becoming dependent on, let alone controlled by, the Chinese state. When Australia had the audacity to suggest an independent investigation into the origins of Covid-19, the PRC blocked Australian exports like barley, beef, wine and coal. Australia surprised the alarmists by successfully compensating through selling more elsewhere[↗], but it had the advantage that production had stayed largely Australian-owned. 

Access to semiconductor technology is vital for technological development, which means economic growth. This may make them essential for regime continuity itself: ‘a matter of survival’, according[↗] to Liu He, the Chinese Vice Premier. Xi Jinping’s ‘China Dream’ of power-parity with the US, or even hegemony, by the 100th anniversary of the PRC in 2049 is not possible without the right semiconductors. Xi Jinping said[↗] in 2016 that ‘our dependence on core technology is the biggest hidden trouble for us’. 

It is possible that Taiwan’s concentration of semiconductor fabs and the high barriers to entry to building leading-edge foundries on the mainland are helping motivate Xi’s preparations for possible invasion of the island. But fabs require very high skills, which makes them vulnerable to workers simply leaving. The PRC will also continue to need foreign manufacturing equipment, while advanced wafers need ultra-pure silicon with contamination no higher than the equivalent of one grain of sand in 16 Olympic swimming pools[↗]. All this makes the industry intensely vulnerable to sanctions or even sabotage. 

Beijing has other motives for invasion that have nothing to do with technology, but Taiwan’s vital position in the semiconductor supply chain means an invasion could cause economic chaos. In the late 1940s[↗], one transistor could be installed on a petal-sized silicon wafer: forty years later that wafer could hold a million transistors, as Moore’s Law described the density of transistors on a chip growing by 100% every two years. The space between two transistors fell from 10 micrometres in 1971 to 5nm in 2020. This is expected to stop at 2nm, where further miniaturisation becomes impossible. Each leap[↗] has led to huge increases in speed: a 200mm wafer featuring 5nm process technology allows 15 times the number of transistors of 30nm. Today, only TSMC and Samsung make 5nm semiconductors, as the East Asian giants increase their technological lead. 

This is not the place to debate the likelihood of a Chinese invasion. But while it remains a possibility, alternative supplies are needed, not increased Chinese control.

Tomorrow, in the second part of this series, Dr Radomir Tylecote will assess the strategic importance of semiconductor technology and manufacturing capabilities to the UK.

Dr Radomir Tylecote is Director, Defence and Security for Democracy, at Civitas. This two-part article is based on his recent report on this area ‘The acquisition of Newport Wafer Fab by China’s Wingtech’, with Henri Rossano. In February 2021 he published ‘Inadvertently Arming China? The Chinese military complex and its potential exploitation of scientific research at UK universities’ with Robert Clark.

Join our mailing list!

Stay informed about the latest articles from Britain’s World

Leave a Reply

Your email address will not be published. Required fields are marked *