ChinaTech - China's Semiconductor Industry
This new segment by Shobhankita Reddy is your go-to newsletter for updates and perspectives on China’s tech ecosystem. This edition studies the government policy instruments and recent developments in China’s semiconductor industry.
This edition is borrowed from work previously done by Saurabh Todi and Pranay Kotasthane at the Takshashila Institution. Find their full research here.
The semiconductor industry is vital to modern society, laying the foundation for most contemporary technology and digital infrastructure. It also has gained significant geopolitical importance as semiconductors are essential for advanced technologies like artificial intelligence, Internet of Things, 5G networks, autonomous vehicles with wide ranging applications in multiple industries - both civilian and military.
This sector has emerged as a key domain of the US-China confrontation. Given the high complexity and resultant costs of chip production, no country is self-sufficient in semiconductors. Instead, semiconductor powers end up specialising in some segments of a value chain that consists of design, materials, fabrication, manufacturing equipment, and outsourced assembly and test (OSAT). A typical semiconductor production process spans more than four countries and three trips around the world, travelling 25,000 miles and spending 12 days in transit. This exemplifies the nature of global value chains prevalent in a complex, interconnected global economy today.
China has long sought to develop indigenous chip-making capabilities, with first efforts dating back to 1956 when semiconductor technology was first identified as a priority area under the ‘Outline for Science and Technology Development, 1956-1967’ leading to the establishment of university programs and state-owned factories focused on R&D and manufacturing of semiconductors. By 1956, a state-run lab had made its first transistor, and by 1965, its first chip, which put China ahead of Taiwan and South Korea. Mao’s policy of self-reliance (自力更生) seemed to be bearing fruit. However, these efforts petered out as state-run firms couldn’t keep pace with their global competitors. Moreover, the cultural revolution squandered even limited gains from this initiative, and by the 1970s, most Chinese factories produced only basic components.
Under Deng Xiaoping’s economic reforms, China once again renewed its efforts to modernise its semiconductor industry by creating a ‘Computer and Large Scale IC Lead Group’. By 1985, state-owned enterprises had imported manufacturing lines, and China attempted to create national champions by concentrating resources in a handful of enterprises. By the late 1980s and early 1990s, China also tried a hybrid model of industrial development, facilitating joint ventures (JVs) with foreign firms, negotiating technology transfers and endowing these firms with additional funding. For example, China attempted to develop Huajing into a leading integrated device manufacturer through a JV with Lucent Technologies (US).
These efforts, too, were unsuccessful. One American researcher visiting a Shanghai factory in the mid-1980s found it was producing chips that were 10–15 years out of date on wafers with yields as low as 20 to 40 per cent.
After the failure of large-scale, state-led projects, China pivoted from state-led approaches to a more hybrid approach, led by the private sector, to innovation. This entailed government support with instruments such as government procurement and tax exemptions for domestic companies operating in fabless design and IC manufacturers. Douglas Fuller, in his book Paper tigers, Hidden Dragons, details that the inflow of FDI led to the rise of hybrid firms (with Chinese management and foreign capital) and foreign-invested enterprises that have helped China develop its semiconductor industry. The most successful such example is the Semiconductor Manufacturing International Corporation (SMIC), which was founded by a Taiwanese veteran of Texas Instruments (US) and Worldwide Semiconductor Manufacturing Company (Taiwan) as a wholly foreign-owned foundry based in Shanghai. Since starting production in 2002, it has emerged as the largest and most advanced chip maker in China. SMIC has developed partnerships with foreign firms and recruited ethnic Chinese engineers (primarily returnees from the United States, Taiwan, and Singapore) to be placed among the top five foundries globally.
In more recent years, since Xi Jinping assumed power, China returned to a more direct state intervention model with a push for indigenisation and self-sufficiency in the sector.
The “Made in China 2025” initiative, unveiled in 2015, sets aspirational goals for China to achieve 70% self-sufficiency in semiconductors. The roadmap specifies that, by 2030, segments of the Chinese IC industry should reach advanced international levels. China’s 13th FYP (2016–20) further prioritises the development of dynamic random-access memory (DRAM) chips to lessen its dependence on memory chips from the United States. In the 14th Five-Year Plan, semiconductors were explicitly mentioned as a strategic technology priority requiring a whole-of-society effort.
Three iterations of the semiconductor ‘Big Fund’ have been set up. Apart from spurring domestic entrepreneurship, the fund serves other purposes - it funds outbound FDI to acquire foreign companies while providing funds to facilitate inbound FDI, such as greenfield investment and JVs with foreign companies in the semiconductor sector. In addition, the government employs a wide range of policy instruments, including tax breaks, grants, equity investments, low-interest loans, discounted land and reduced utility rates for the industry, all amounting to over a hundred billion dollars of public expenditure. China's Ministry of Industry and Information Technology has urged Chinese EV carmakers to procure 25% of their chips from domestic sources by 2025, even though they are costlier than foreign alternatives by up to 30% - with implications for market distortion while ensuring a market for locally produced chips.
Despite these sustained efforts via multiple national initiatives and policy instruments through the years, analysing the Chinese semiconductor industry reveals a mixed bag.
China today is a world leader in ATP capacity and supplies more than 38% of global demand, as of 2021, second only to Taiwan. China also dominates in the production of minerals and industrial raw materials necessary for chip manufacturing. However, in semiconductor fabrication, part of the value chain with the most government intervention, Chinese firms have had limited catch-up. Chinese fabs can produce lower-end logic and analogue chips for consumer, communications and industrial end markets, as well as higher-end chips with low yields. China is a world leader in wafer fabrication for logic chips at or older than 28nm. In fact, the projected production of these chips over the next decade, according to the Semiconductor Industry Association, is far likely to outstrip global demand, with much of this being produced in large fabs in China. This has the potential to be yet another example of Chinese overcapacity causing imbalances and pushing global prices down.
In the design segment, there are now over 3000 companies in China, with double-digit annual revenue growth, with various manufacturers beginning to establish fabless design units and together accounting for around 20% of the global fabless market. However, as evidenced by the quality of indigenous chips, the success of Chinese semiconductor design has been checkered. Chinese firms have focused on less advanced consumer electronics, industrial control systems, and intelligent device chips, but are less competitive in more advanced CPUs, GPUs, corresponding high-end servers and power management.
It is noteworthy that despite a significant portion of registered capital being directly or indirectly controlled by the state, the semiconductor industry segments that have succeeded have done so with less government control. The widespread corruption, mismanagement and inefficiencies in the Big Funds have been detailed previously. As Douglas Fuller puts it in his book, Paper Tigers and Hidden Dragons, Hybrid firms have, indeed, been the ‘hidden dragons’ of the Chinese economy.
And yet, China continues to remain massively dependent on the global market for its domestic needs. In 2023, for example, 29% of all global semiconductor sales went to China, the highest of any market. Less than 30% of China’s domestic needs are serviced by its firms, a significant gap as compared to the 70% target stipulated by the MIC initiative. Chinese firms only service 7% of the global semiconductor sales, and as indicated by their low reinvestments in R&D at 7.6% compared to their peers in Europe, Japan, Taiwan, South Korea as well as the US (where the reinvestment rate is 19.5%), they also have weaker business models and balance sheets.
China is heavily dependent on foreign firms for EDA software and semiconductor manufacturing equipment. This is a critical hindrance in China’s path to self-sufficiency. While China now accounts for 20% of global equipment spending and 18% of global equipment imports, this is in jeopardy in light of the potential coordinated export controls from the US, Japan, Netherlands, and other countries.
More generally, China’s dependence on foreign firms for its semiconductor needs, despite persistent attempts at import substitution that have only yielded progress in some parts of the value chain and specific types of chips, highlights the challenges it faces for technological self-reliance. A more fragmented and less collaborative global environment will further rein in China’s ability to absorb technology to advance itself over the long term.