Great Leap Nowhere: The Challenges of China’s Semiconductor Industry
China is struggling in the battle for advanced semiconductor technology. With President Joe Biden’s most recent round of export controls on semiconductors, China is now facing an increasingly urgent challenge as it seeks to ramp up its domestic innovative capacity for high-end chips. Yet at the turn of the new year, Bloomberg reported that China is pausing its investment in its domestic semiconductor industry, most likely due to financial strain from COVID-19, ineffective subsidies, and corruption. This came only two weeks after the news that Beijing is preparing a one-trillion yuan ($145.61 billion) incentive package to beef up the competitiveness of its indigenous chip industry.
While the gigantic cost of the COVID-19 lockdown necessitates some belt-tightening, Beijing’s rollback of its massive chip investment should not be surprising. It is no secret that China’s previous efforts to improve its chip innovation did not bear much fruit. The reason for the lack of success, however, is more than the technical challenge of independently developing one of the most sophisticated products in the world. Rather, the political and institutional factors in China’s science and technology innovation base are also impediments to its desired success in reaching a chip breakthrough. Not only does China’s top-down model of innovation overshadow its domestic ecosystem of chip development, but the absence of effective oversight and the perverse incentives from the indiscrete dispersion of local government funding together render prospects for chip innovation grim.
For U.S. policymakers, the messiness of China’s domestic semiconductor industry should be a significant factor to consider when devising export controls and enhancing cybersecurity infrastructures for both domestic firms and key allies. As the constraints of China’s innovative potential are intrinsic to its political system, the United States and its allies can be cautiously optimistic about their long-term competitiveness in semiconductor technologies.
Lots of Investment, Little Return
The Chinese government has allocated the semiconductor industry not only one trillion yuan through state capital such as the Integrated Circuit Investment Fund (the Big Fund) but also high political priority, directing both political efforts and the market to infuse the chip industry with resources. Since the Third Party Plenum of 2013, the government has made semiconductors the country’s top industrial innovative priority. The “Made in China 2025” campaign was launched in 2015 by Premier Li Keqiang to target China’s reliance on foreign technology and enable globally competitive indigenous innovation. This represents a significant political and economic move to enhance China’s innovation base. Dozens of companies that are relevant to semiconductor design and manufacturing have received favorable government policy and capital to expand foundries, hire talent, and build facilities on land acquired with substantial support from local governments.
These efforts, however, did not seem to move China up in the semiconductor value chain. Even after billions were thrown at the problem, indigenous production is far from being a reality. Despite some progress in independent chip design for a variety of products ranging from cloud computing to smartphones, the country still could not break free from the foreign-dominated supply and manufacturing chain. Huawei’s cutting-edge Kirin 9000-series chip, which uses a 5 nanometer process, now faces extinction as Huawei became the target of a series of devastating sanctions that targeted China’s inability to independently build the necessary electronic design automation and nanometer lithography machines, necessary ingredients to turn these advanced chip designs into products.
Political and Institutional Constraints
This frenzy of state capital investment, political salience, and the nationalist narrative sounds eerily similar to the disastrous Great Leap Forward of 1957, which was about increasing steel production. Unfortunately, this 21st-century “Great Leap Forward” carries over some of the persistent political and institutional problems that underpinned the 1957 movement and prevented that leap from going forward.
First, the government does not seem to direct investment based on the requirements for independently creating essential chip manufacturing technologies. The top-down process by which industrial policies are made is often driven by political priorities, not scientific realities. Today’s outsized investment into the chip industry seldom goes to the fundamental technologies that underpin the advanced chip supply chain. To break through, China must invest heavily in the basic science that conceived these technologies, not just those that built or fabricated them. Yet developing basic science to achieve original innovative breakthroughs entails substantial sunk costs and long product development cycles that Chinese political actors currently are not incentivized to commit to. Massive government programs like the Big Fund often distribute funding expecting expeditious and demonstrable results that are tied to promotion decisions and future funding. As such, investing in basic science developments with long trial-and-error cycles becomes a risky proposition that disincentivizes both tech companies and officials.
In the case of lithography, for instance — carving billions of transistors on a piece of silicon at a microscopic level using ultraviolet light — the technology that makes it possible is beyond engineering and requires cutting-edge achievements in material science, laser physics, and chemistry. An essential element in the lithography process, known as photoresist, is a highly sophisticated photosensitive material that was invented in the 1920s and perfected throughout the 20th century by multiple American, Japanese, and European scientists. Photoresist, except for limited low-end products, is among the key technologies that China needs to import, as Japan and the United States have near-exclusive control. Thus without investing substantial time and sunk cost in fundamental sciences, China will likely remain dependent on imported materials.
China’s initial phase of investment since 2014 did not target these fundamental areas, however. Out of the 98.72 billion yuan raised by Big Fund, 30 billion went to expanding foundries and 20 billion went to integrated circuit design, both of which still rely on readily available Western technologies to move forward. Only a fraction of the investment went into equipment and material. According to an itemized list of investments from the first phase of the Big Fund, only about 1.3 billion yuan went to equipment companies and about 1.4 billion yuan went to companies that work on material science. Similarly, in March 2020, a more comprehensive investigation using China’s data investigation platform Tianyancha (天眼查) found that only 6 percent of the Big Fund I went to material and equipment but 70 percent went to fabrication, which still extensively relies on foreign technologies for high-end products. The second wave of investment starting in 2019 did not seem to value basic research either. It invested in the integration of upstream suppliers and downstream applications, usually in the form of connecting regions of the semiconductor industry (such as the Yangtze River Economic Belt) and industrial parks. In sum, the Big Fund was primarily devoted to the applied and production side of the semiconductor as opposed to the basic science side that would hold the key to original innovation and indigenous breakthroughs.
Secondly, and relatedly, the massive injection of capital was done with little institutional oversight but substantial political expediency, creating incentives for companies and local governments to jump on the bandwagon regardless of their capabilities. Contrary to what is expected of a centralised authoritarian state, provincial and local governments in China have substantial discretion over government debt and spending, which can negatively affect science and technology funding and innovation. Local officials, politically incentivized to win favors from superiors and pad performance records for promotion, hand out funding and contracts to entities that claim to be able to make breakthroughs in semiconductors. This has resulted in a perverse incentive for businesses to rip off local government funding, leading to rampant corruption and questionable investment decisions such as funding companies with no semiconductor experience to work on chips. Due to the financial attraction of the Big Fund, from January to October 2022 there appeared more than 58,000 new integrated circuit-related firms, over 13,000 of which were moved from their previous operations in unrelated sectors (a practice known as zhuanchan转产 in Chinese). According to the spokesperson of the National Development and Reform Commission, the industry had a significant “three-nos” problem — companies with no experience in semiconductors, no technical know-how, and no talent.
Without any independent oversight, the decision of many profit-driven firms to jump on the semiconductor bandwagon unsurprisingly led to fraud, corruption, and misallocation. The Chinese semiconductor industry has long suffered from these issues as local governments were giving out subsidies and funding with little due diligence. In the 2003 Hanxin fraud, a scientist named Chen Jin raked in over 100 million yuan of government subsidy, an amount greater than several years of investment in the semiconductor industry at that time, by re-packaging Motorola chips that used a then-cutting-edge 180 nanometer process and claiming them as his own design. In September 2018, the Hongxin Semiconductor Industrial Park was launched in Wuhan under the personal watch of Mayor Zhou Xianwang. The initial capital investment of 2 billion yuan was done by two entities backed by state capital. For two years, Hongxin became the most invested program in Wuhan, having secured 128 billion yuan from the Wuhan Reform and Development Commission. The program later proved to be a complete failure as the founders had no knowledge of chips whatsoever, and the facilities were abandoned. Around the same time, similar stories emerged in Hebei and Nanjing, with multi-billion-yuan chip programs fraudulently making big promises to siphon off government funding. By mid-2022, when the Chinese Communist Party began investigating the egregious practices in the chip industry, billions had been wasted in the frenzy.
Can China Leap Forward?
Even if the anti-corruption investigations manage to cleanse the industry of opportunistic and unscrupulous businesspeople, China still must confront the same problem that it encountered before— achieving significant breakthroughs using a backward and statist innovation ecosystem. Two particular challenges stand in China’s way.
First, the technology involved in the semiconductor industry is exceedingly complex and requires original innovation, rather than duplication or imitation, to move forward. This is particularly problematic because China still uses a top-down approach to innovation that has been found to impede originality and “breakthroughs in both new component technology and architecture.” Some of China’s most promising semiconductor companies demonstrate why. The Yangtze Memory Technologies Corporation was, as the Biden administration’s 100-day Chip Supply Chain Report described it, China’s “national champion memory chip producer.” Yet, in less than a year, the state-owned Assets Supervision and Administration Commission of Wuhan acquired the controlling share of the corporation through an investment firm and ordered it to supply Huawei and Hikvision, earning the six-year-old company a place on the U.S. 2022 Entity List. Recent anecdotal evidence suggests that this proved to be a fatal blow to the company’s upward potential, as it began laying off employees and asking them to pay back the price difference of the houses they bought with a discount while employed.
Second, the goal of semiconductor self-sufficiency is too big for any one country, no matter how advanced. The semiconductor supply chain is not a singular effort. Steppers, photoresists, and electronic design automation from the United States, the Netherlands, Japan, and Taiwan are all indispensable in the global semiconductor industry. In other words, it would be unrealistic even for the United States to design and produce advanced chips without the help of its allies who hold some degree of monopoly over specific necessary technologies for high-end chip production. China has made some progress in independently securing the technology for the last phase of chip production: assembly, testing, and packaging. But with many key fundamental technologies under-invested, there remains a long way to go for China to liberate itself from the technological chokepoints and to reach the self-sufficiency envisioned by Chinese leaders.
In these circumstances, what innovative potential does China have for chips? The answer depends on how radical or original the innovation is. China has a great track record of creative adaptation or “re-innovation,” such as using older platforms to incrementally improve a certain part of the technological development chain. The Semiconductor Manufacturing International Corporation (SMIC), China’s largest chip company, achieved a breakthrough in the quasi-7 nanometer process using old deep ultraviolet tooling acquired from the Dutch. This is consistent with China’s history of being very skilled at making incremental improvements to legacy technologies such as ballistic missiles. But these lower forms of innovation frequently encounter bottlenecks as they are band-aid solutions to the underlying problem of having an inadequate reserve of fundamental sciences and technologies. Experts have questioned whether the Semiconductor Manufacturing International Corporation’s breakthrough can be sustained or scaled to production due to the low yield and high cost of using older deep ultraviolet tooling for commercial chips.
All that said, while the Big Fund is on pause, China could move to pour money into the fundamental sciences. But doing so necessitates changes in how the political actors are incentivized and how research and development institutions are managed to prevent fraud and sustain costly explorations. This systemic overhaul of the political aspect of the science and technology ecosystem would directly affect the current state of government-business relations and thus would likely prove prohibitively slow and difficult.
Implications for the United States
The U.S. approach to maintaining its semiconductor technological dominance over China with its allies should be informed by the domestic institutional constraints of China’s semiconductor industry and its innovation model. Limiting factors such as statist intervention, lack of oversight, and overall top-down innovation are inherent to the Chinese political system and thus are likely to persist. While this is not to say that the Chinese chip industry is not capable of innovating at all, we should expect more incremental improvements with immediate bottlenecks but fewer game-changing radical innovations.
At the same time, U.S. policymakers should be keenly aware that China’s relative success with creative adaptation means that it can boost certain sectors of the chip industry by exploiting leaky export controls and engaging in cyber espionage. As seen in the case of Semiconductor Manufacturing International Corporation, China could still try to work around certain technological hurdles with already acquired foreign technology and funnel the profit into areas in which they are currently lacking. This means that the U.S. government should continue to monitor Chinese imitation-style innovations, and work closely with key semiconductor allies to make the export controls against China as airtight as possible. This may entail nuanced multilateral negotiations to address the specific interests of key allies like the Netherlands, which is not fully aligned with the U.S. position on the bans on older machines. Additionally, the United States should consider supporting the cybersecurity efforts of its domestic semiconductor firms to mitigate the risk of cyber espionage, which China extensively employs to lower the hurdle of original innovation both in military and civilian settings.
Elliot Ji is a Ph.D. Candidate in international politics at Princeton University. He is a 2023 Nuclear Scholar of the Center of Strategic and International Studies’s Project on Nuclear Issues and a 2021 recipient of the Paul & Daisy Soros Fellowship for New Americans.
Image: Wikimedia Commons