Chinese chip making: is this the Dreadnought moment?

China is making efforts to build a chipmaking industry of its own. Can it be done?

Futuristic circuit board

It’s widely understood that a lot of the technology business relies on Chinese manufacturing. From the data centre to the network hub to the cloud edge to the personal computer or mobile device or Internet-of-Things endpoint, hardware is likely to have “Made in China” written on it.

What’s less well known, however, is that if you open up such a piece of kit, the chips inside it are likely to have been imported to China from other nations. China supplied just 19 per cent of its own semiconductor demand as of 2019, according to IC Insights. Even if a given chip was made in a Chinese factory, most of the key machinery in that chip factory will have been imported because there are no Chinese suppliers able to build such machines.

This dependence of Chinese manufacturing on foreign supplies of semiconductors and the machines needed to make them makes China far weaker than it seems in the ongoing trade war with the US. Chinese networking giant ZTE almost collapsed in 2018 when supplies of US chips were cut off. Huawei has faced similar issues.

But chipmaking technology is changing radically. Such changes can sometimes have the effect of invalidating a lead built up over a long period and effectively start a new race from scratch. An historical example comes from the turn of the last century, when the British navy was more than twice as powerful as the combined fleets of the next two most important naval powers at that time, Germany and France.

But then the first Dreadnought-class all-big-gun battleship was launched, in 1906. All previous types of warship effectively became obsolete overnight: the only thing that mattered in naval power was the number of dreadnoughts. The Royal Navy remained the most powerful in the world for some time thereafter, but now it needed to bring all its force to bear to overcome just one serious opponent: it could no longer take on two or three at once and still win easily.

Chipmaking today is passing through a similarly major shift. Moore's Law remains on the books, but the process which drives it has changed. The ‘Law’ states that the number of transistors in an integrated-circuit chip of a given area doubles every 18 to 24 months, a dictum laid down by Intel co-founder Gordon Moore in the 1960s. For more than four decades the industrial process behind the ‘law’ was the shrinking size of transistors and circuits built on the surface of a silicon wafer.

To continue reading this article register now