CNA file photo

By Alison Hsiao, CNA staff reporter

On his last day as chairman of Taiwan Semiconductor Manufacturing Co. (TSMC) on June 4, Mark Liu (劉德音) said it would be “impossible” for China’s Huawei to catch up to TSMC in semiconductor development.

The blunt comment at TSMC’s shareholders meeting renewed debate in Taiwan on China’s prospects of making leaps in technology on its own despite facing American sanctions that began in 2019.

One skeptic in the debate is Lin Hung-wen (林宏文), a veteran journalist and analyst who has followed TSMC’s evolution over the past 30 years.

In a recent interview with CNA, he questioned whether China’s attempt to build a home-grown semiconductor industry through the injection of massive amounts of money would succeed.

In his book on Taiwan’s chip industry and TSMC published last year, Lin noted that Huawei had the fourth highest research and development (R&D) spending of any company in the world in 2022, trailing only Alphabet (Google’s parent company), Meta, and Microsoft.

More noteworthy, however, is that its US$23.8 billion R&D spending accounted for 26 percent of its US$91.5 billion revenue, which “showed its ambition,” Lin said.

“This R&D to revenue ratio is [unusually high], especially when compared to other major tech companies whose R&D spending accounts for 5-8 percent of their revenues,” he said, citing TSMC’s 7.2 percent ratio in 2022.

Lin was doubtful, however, that Huawei could rely on Chinese chip manufacturers to back further breakthroughs in 5G telecommunications and electronics.

Money-pumping strategy

When Huawei came out with its Mate 60 Pro phone, equipped with an advanced 7-nanometer chip made by Shanghai-based Semiconductor Manufacturing International Corp. (SMIC), in August 2023, it stunned the industry.

Huawei’s fabless IC design company Hisilicon had been cut off from its access to TSMC chips since 2020 and had since turned to Chinese chipmakers, leading some to believe that China’s domestic semiconductor sector and SMIC had cracked the code and were catching up to TSMC more quickly than expected.

It was later reported, however, that the breakthrough used technology from U.S. suppliers, illustrating how hard it is for China to move into highly advanced chip manufacturing on its own.

“There is simply no one country that can do it alone,” Lin said, because the industry has become increasingly dependent on a division of labor, a trend that even massive spending by Beijing will be hard-pressed to reverse.

In May, Beijing injected an additional US$47.5 billion into the China Integrated Circuit Industry Investment Fund — initiated in 2014 to help China achieve semiconductor self-sufficiency — after putting up US$47.4 billion in funding during its first two phases.

Lin said every step in the manufacturing process is highly sophisticated or requires a supporting ecosystem, such as the Netherlands’ ability to make high precision extreme ultraviolet (EUV) lithography machines or the U.S. setting chip design specifications.

That global division of labor is so firmly embedded in the industry that no single country’s success can be “comprehensive,” even for an economy with pockets as deep as China’s, Lin added.

History is also not on China’s side, Lin said, citing the work of Yasheng Huang, a professor of global economics and management at the MIT Sloan School of Management.

Huang’s research has found that when China has made technological advances in the past, even during its many dynasties, it has been open to the world or allowed a diversity of ideas.

More recently, Lin said, China “benefited greatly from the world’s globalization and liberalization, and its enterprises heavily depended on Hong Kong as a liberal financial center with the rule of law to raise funding.”

That openness has since waned, and with Hong Kong’s downfall as a liberal and open gateway, China faces serious challenges in continuing its economic success and chip development, Lin argued.

Lin acknowledged that the final chapter has yet to be written on China’s push to be self-sufficient in advanced chips, but he could not see how it could succeed.

“The advanced processes simply require state-of-the-art technologies and equipment that are not really possible by self-reliance,” he said.