If Tech Transfers Didn’t Build China’s Chip Industry, What Did?
Last week, American President Joe Biden signed the CHIPS and Science Act into law, setting the stage for the United States government to hand out $52.7 billion in subsidies to chip manufacturers. The bill, which was presented by the White House as a way to “counter” China in the global race for advanced computing technology, bans any firms receiving funds under the Act from investing in advanced chip production capacity in China for a period of 10 years.
The underlying rationale for this provision can be traced at least in part to a decade-old myth surrounding the practice of “forced technology transfers.” Two influential White House reports produced and released by the Donald Trump administration lay out the argument in detail, accusing the Chinese government of leveraging market access to compel foreign companies to transfer advanced technologies to Chinese firms and claiming that the rise of China’s tech industry has been driven by these forced technology transfers. This, the Trump White House argued, justifies the use of export controls to cut off the access of leading Chinese firms to advanced American technologies.
Four years into the trade war, China’s high-tech sector remains resilient under the sanctions. Yet the CHIPS act suggests the myth of forced technology transfers as fundamental to Chinese growth continues to haunt U.S. policymaking. But in my new book, “China’s Drive for the Technology Frontier: Indigenous Innovation in the High-tech Industry,” I argue that, compared to the spotty record of China’s technology transfer policy, the real driver of China’s technological progress has been indigenous innovation, a process in which domestic firms innovate to solve problems and create products for the local market based on a mix of knowledge actively learned from advanced countries as well as from China’s own science and technology base.
Advanced semiconductors, or chips, are a critical technology and a “chokepoint” for the Chinese economy in the trade war. Although China’s history of fabricating semiconductors dates back to the 1960s, it was only after the 1978 economic reforms that the Chinese government made a concerted effort to establish a modern chip industry. In the more than four decades since, the Chinese chip industry has been through three phases of development: a period emphasizing foreign technology transfers (from the late 1970s to 1999); a period of globalization and embedding in global supply chains (2000-2014); and the current phase, characterized by the rise of indigenous innovation.
In the first phase of development, which most closely resembles the “forced technology transfer” stereotype, the Chinese government implemented an industry policy explicitly centered around “trading market for technology,” or TMFT. As the term suggests, the policy required foreign companies to introduce advanced technologies to China as a precondition for access to the vast and growing Chinese market. Deployed to great success in the automobile, telecommunications, and electronics industries, the TMFT requirement could be fulfilled through transferring complete sets of equipment or blueprints, training staff, and forming joint ventures with state-owned enterprises (SOEs).
In the chip industry, the key beneficiaries included two state-owned enterprises and four Sino-foreign joint ventures. While the operations of these firms laid a solid foundation for the future, it is important to note that none of them became the kind of globally competitive enterprise imagined by the Chinese state. Likewise, two of the most expensive state-sponsored technology transfer projects, Project 908 and Project 909, failed in their aim of fostering a vertically integrated “national champion” capable of competing in the global chip industry.
There were a variety of reasons the TMFT failed to achieve the desired results. To start, foreign partners lacked the incentives to transfer truly critical knowledge to a potential competitor, and thus core technologies were often inaccessible to Chinese managers and engineers. More critically, the SOEs at the time lacked strategic control, such as authority and capacity over crucial investment decisions, meaning they had little say in what technologies were to be transferred in the state’s negotiation with foreign firms.
For example, a manufacturer of discrete semiconductors for consumer electronics was tasked by Project 908 with using technology transfers from American telecommunications giant AT&T Lucent to make advanced communication chips. This type of gap between a firm with a weak technology base and the government’s ambitious technology transfer plans often led Chinese managers to cede practical control over projects to foreign partners with better knowledge of technology, management, and markets.
The establishment of Semiconductor Manufacturing International Corporation (SMIC) in 2000 marked the advent of the globalization era in China’s chip industry. Founded by a team of ethnic Chinese and overseas returnees with experience in Silicon Valley and Hsinchu, and with funds raised from leading venture capitalists and the Chinese government, SMIC brought to China the pure-play foundry business model, pioneered by the Taiwan Semiconductor Manufacturing Company (TSMC).
Separating extremely capital-intensive foundries from chip design, independent foundries like SMIC and TSMC serve a broad base of customers, from semiconductor design houses to integrated device manufacturers.
With orders and technology licensed from its overseas customers, as well as support from various levels of the Chinese government, SMIC quickly began catching up with industry leaders. Importantly, unlike the national champions of the previous era, SMIC had the freedom to pursue an independent strategy, which it used to build up massive economies of scale in foundries, seek supply chain partners, and compete for a share of the global market. In the mid-2000s, SMIC was briefly the world’s third largest pure-play foundry after TSMC and the United Microelectronics Corporation.
SMIC’s early success was built on embedding itself in the global value chain, but it soon ran into trouble when TSMC sued it for patent infringement and the theft of trade secrets in U.S. court. As part of the suit, SMIC exports to the United States — the company’s largest market — were blocked twice, in 2002 and 2006. Although the suit was eventually settled, the company’s growth never recovered.
Insufficient “patient” capital is also partly to blame. Chip foundries are notoriously capital intensive, and SMIC’s capital expenditures, a critical indicator of investment in production and technology, averaged $700 million between 2005 and 2013. For context, TSMC’s capital expenditure grew from $2 billion to $10 billion annually during the same period. By the beginning of the 2010s, the Chinese chip industry had lost momentum, and the production and technology gap between the leading firms in China and the rest of the world had widened.
The next phase of the Chinese chip industry wouldn’t begin until the establishment of the National Integrated Circuit Industry Investment Fund — commonly known as the Big Fund — in 2014. With an eye toward solving the industry’s capital problems, the Big Fund mobilized a large amount of capital for intensive investment. By the end of its first investment phase in 2018, the Big Fund had invested upwards of 100 billion yuan in the chip value chain, according to estimates, covering everything from design and fabrication to materials and equipment.
A major mandate of the Big Fund is to strengthen the weak links in the domestic semiconductor industry. The fund’s initial focus was chip fabrication. More recently, it has shifted investment to semiconductor equipment and materials as these fields have emerged as new chokepoints in Sino-American competition.
A defining characteristic of the indigenous innovation phase is not simply making chips smaller, but the innovation of higher-quality, lower-cost chips for the local market. By filling these niches, indigenous firms have been able to build profitable business capable of investing in the R&D needed to one day become industry leaders.
Two startups, the Yangtze Memory Technology Co. (YMTC) and ChangXin Memory Technologies (CXMT), have made breakthroughs in the indigenous production of memory chips, a market segment accounting for a third of China’s $400 billion in chip imports in 2021. YMTC has leveraged its innovative Xtacking technology to earn about a 5% share of worldwide NAND chip production, while CXMT has built on the technology and patent portfolio acquired from Qimonda AG, a bankrupted German DRAM maker, to manufacture DDR4 DRAM chips.
Although no longer reliant on forced technology transfers, the Chinese chip industry still depends on returnees and overseas talent for their engineering and management expertise, as well as technologies licensed or bought abroad. This has posed challenges: By 2019, SMIC made a breakthrough into 14-nanometer FinFET technology through indigenous research and development. Yet, U.S. government restrictions on the sale of advanced EUV lithography machines mean there are significant technological obstacles to SMIC developing cutting-edge sub-10 nanometer technologies.
Meanwhile, as recent corruption investigations indicate, large state funds are not always efficient allocators of capital. The Big Fund, which before the recent scandal was seen as conservative in its funding decisions, differs radically from the traditional East Asian model followed elsewhere, which relies on big banks and industrial conglomerates to channel funding into the semiconductor industry.
Nevertheless, while the U.S.-China trade war that broke out in 2018 has slowed China’s technological development in some spheres, it has also accelerated the process of indigenous innovation. Ultimately, this drive, rather than any sanctions, will determine the future of China’s chip industry.
Editors: Cai Yiwen and Kilian O’Donnell; portrait artist: Wang Zhenhao.
(Header image: Ingram Image/VCG)