China’s public and private sectors are trying to accelerate the country’s domestic chip-making capabilities, but enthusiasm doesn’t always yield efficiency.
A hasty push for independence in core techChina imported USD 350 billion worth of semiconductors in 2020, according to the China Semiconductor Industry Association (CSIA), up from around USD 300 billion in 2019, when the size of semiconductor imports already eclipsed the value of the country’s oil imports. To reduce its reliance on semiconductor imports, the Chinese central government is providing funds to local governments to invest in chip production. Subsidies and investment from local governments and the private sector into China’s domestic chip companies from 2014 to date amount to at least USD 170 billion, according to the China Securities Journal. The investment momentum is only picking up, as Chinese chip firms raked in a total of RMB 227.6 billion (USD 35.2 billion) via the primary and secondary markets in 2020, up 407% year-on-year (YoY) compared to 2019. “The domestic chip market is currently overheated. Some semiconductor projects’ valuations have reached billions of yuan by the angel or Series A round, so there must be a bubble. It is still too early to know if this bubble is good or bad for the industry as a whole,” Song Chunyu, a partner at Lenovo Ventures, told 36Kr. Developing a chip industry from scratch is inherently a capital-intensive business. It takes USD 15–20 billion to build a foundry that may be obsolete five years later. A degree of speculation is perhaps to be expected, but following the recent flurry of activity and capital in the Chinese chip sector, some projects simply did not possess the technical aptitude to make products like semiconductors, while others were just fraudulent initiatives.
More investors than quality projectsDespite the unbridled enthusiasm from investors looking to ride the macroeconomic wave of China’s emphasis on domestic chip production, choosing the right startup to back can be tough. “Investing in semiconductors can be difficult because of the long industrial chain and complicated processes involved. The cost of tape-out alone can reach several million dollars, including significant labor expenses. A competent chip engineer needs to be trained for at least five years, which costs around USD 1 million, while it takes at least 18 months for a competent team to build a chip,” Yang Lei, a partner at Northern Light Venture Capital, told 36Kr. In some cases, investing in the wrong project can be truly disastrous. One of the most egregious failed projects was the HSMC debacle in 2020. Cao Shan, an individual with no more than primary school education, managed to dupe investors, including the Wuhan government, into committing RMB 15.3 billion (USD 2.4 billion) towards building a high-tech semiconductor production company that never materialized. While the scandal hogged headlines, it was hardly the first, and certainly not the last Chinese chip startup, to go bust. In late 2015, the Nanjing local government was one of the first provincial administrations keen to adhere to national directives and incubate a domestic chip champion. City officials chose to participate in a USD 2.5 billion investment round for a newly founded local startup named Nanjing Dekema Semiconductor Technology. The firm had ambitious goals—aiming to deliver 40,000 wafers per month—yet the company never got off the ground. Under the leadership of the company’s founder Li Ruiwei, who had minimal experience in the semiconductor field, the startup floundered and eventually claimed bankruptcy in late 2019, having been unable to pay employees, taxes, and other financial commitments. Li was a notorious speculator who had been involved in starting two other chip projects in Shanghai and Ningbo, both of which remain unfinished. The construction of Nanjing Dekema’s production facilities halted around 2018, while the company never produced a single chip and is now mired in lawsuits from former employees, suppliers, and contractors.
A similar situation unfolded in China’s Shaanxi Province. Shaanxi Kuntong Semiconductor Technology was founded in October 2018 with a total planned investment value of RMB 40 billion (USD 6.25 billion), backed by both the Fengxi local government and private capital. The company planned to build a fabrication plant that could churn out 30,000 sixth-generation chips a month. The Fenxi local government, which held a 5% stake in Kuntong, was excited at the company’s potential to generate RMB 26 billion in annual sales while also employing around 5,000 people once the plant would have reached full production capacity.
Yet, this rosy vision for Kuntong never came to fruition. In October 2019, a group of senior executives abruptly resigned from the company, citing the firm’s lack of progress as a driving factor in their departure. The exodus interrupted the startup’s development and normal operations, and by January 2020, employees complained of unpaid wages, with no response from Kuntong. After examining the company’s shareholding structure, it was clear that the man behind Kuntong, Li Xiaoyan, had no prior experience investing in or operating semiconductor companies, exhibiting what is becoming a recurring theme for failed chip projects in China.
Following industry leaders
Despite some high-profile missteps like HSMC, there are also success stories within China’s semiconductor push. China’s leading domestic chipmaker SMIC raised RMB 53.2 billion (USD 8.32 billion) in a historic IPO on the Shanghai Star Market in June 2020 and posted record revenue of USD 1.1 billion in the third quarter of 2020, before finding itself on a US Entity List in December 2020.
However, the Shanghai-based firm managed to continue its revenue growth even after falling afoul of US sanctions, reaching a monthly production capacity of 520,750 8-inch equivalent wafers, which are used in products like mobile phone chips, camera sensors, wireless communication chips, and automobile chips. The company also announced plans to open a new 300 mm fabrication plant in Beijing that will focus on 12-inch silicon wafers, with plans to produce 135,000 pieces per month. Twelve-inch wafers are more advanced than 8-inch devices, but have similar use cases.
SMIC will also open a new production plant in the southern Chinese city of Shenzhen, long the country’s capital of hardware manufacturing. The new facility will cost around USD 2.35 billion, jointly invested by SMIC and the Shenzhen government, and will focus on 28 nm integrated circuits, with aims to produce 40,000 wafers per month. Generally, 28 nm chips are used in a wide range of applications such as central processing units, graphic processors, networking chips, smartphones, tablets, and internet-of-things devices.
With the current global chip shortage set to persist into 2023, expanded production from countries like China can be a welcome supplement to unrelenting global demand.
Younger chip startups such as Yangtze Memory and Changxin Memory are held up as bastions of innovation in the Chinese chip sector. Yangtze Memory, which was only founded in 2016 in Wuhan, is already mass-producing state-of-the-art 64-layer and 128-layer NAND flash memory chips, which are used in several types of storage devices, including SSDs, USB flash drives, and SD cards. The company plans to double its output to 100,000 units per month by the second half of 2021, which will represent about 7% of the global output.
Meanwhile, Hefei-based Changxin Memory specializes in dynamic random access memory (DRAM)—which is used in personal computers, workstations, and servers—and has already captured around 3% of the global DRAM market, currently dominated by Samsung. The company got its start in 2016 via initial investment from the Hefei Industrial Investment Fund.
“Both companies have hired many Korean engineers, who make up most of the global talent in memory chips. Most in the industry today expect that Yangtze Memory will become a significant global player in three to five years,” Dan Wang, a tech analyst at Gavekal Dragonomics, told the Financial Times.
Yangtze Memory’s Wuhan facilities. Source: Yangtze Memory website.
The success of these companies has only emboldened investors in the semiconductor space. Chip-related companies make up nearly 25% of the firms listed on Shanghai’s Star Market, China’s answer to Nasdaq, and in February 2021 alone, 33 semiconductor companies raised more than RMB 5.3 billion (USD 828.5 million) in financing.
Chinese smartphone giants like Xiaomi and Huawei are also directing their investment funds towards the chip space. Xiaomi alone backed 30 chip-related startups in 2020, while Huawei invested in 20 chip projects during the same year.
The market remains promising. Total sales of Chinese-made chips in 2020 reached RMB 884.8 billion (USD 138.4 billion), up 20% YoY, growing at around three times the global average, per data from the CSIA. Though progress is being made in China, US-made chip companies still lead; they generated USD 208 billion in 2020.
Yet, the Chinese investment wave in semiconductor projects is unlikely to abate any time soon. The sector’s future might be bright as long as investors and entrepreneurs can minimize the number of projects that end up like HSMC, Nanjing Dekema, and Shaanxi Kuntong.