How Taiwan won the semiconductor race
- November 3, 2022
- Michael Reilly
This small island state in the northern Pacific dominates the global chip industry — what explains Taiwan's success?
The sometimes critical knock-on effects in so many industries of the shortages of integrated circuit chips, or semiconductors, in the aftermath of the Covid pandemic has thrust into sharp focus the global economic role of a small island state in the northern Pacific. Semiconductors — electricity-conducting chips found in thousands of products today from computers and smart phones to medical equipment and military hardware — have been likened to oil in the 1970s for their crucial contribution to global output, but there is one big difference. Global semiconductor output is dominated by Taiwanese companies, with a single firm, Taiwan Semiconductor Manufacturing Company (TSMC), responsible for around 50 per cent of global semiconductor chip production. By contrast, no single country accounts for more than 18.5 per cent of global oil production. Just as the oil shocks of the 1970s prompted a wave of new exploration, so the belated awareness of this dependency on Taiwan is prompting the US and other countries to invest heavily in their own production of chips.
Given that until the early 1960s Taiwan was a predominantly agricultural economy, with sugar its biggest export, and that at the end of the 1970s the US dominated global semiconductor production with 65 per cent of the total market, policy makers are inevitably now asking how Taiwan’s current market domination was achieved. It is a question answered by both free-market economists and those favouring central planning supply but, in neither case, do these explanations (or theories) provide a full picture.
Yet there has been nothing secretive about Taiwan’s success. Apart from an element of good fortune and good timing — essential to success in almost any field, in broad terms it can be attributed to three factors: one almost universal; one common to much of East Asia; and one perhaps unique to Taiwan.
The near universal factor has been the impact of globalisation, especially in the electronics industry. From the 1970s, American domination of the sector through then-household names such as IBM, Texas Instruments, and RCA, started to be challenged by Japanese competitors. By the mid-1980s Japanese global market share of semiconductor chips overtook that of the US. In response, American companies, together with Dutch firm Philips, at the time also a leader in semiconductor production, started to invest in East Asian countries to take advantage of lower labour costs and skilled workforces. Initially, the new plants were largely limited to final assembly operations of consumer electronics, whose small size relative to value meant the cost of shipping was only a small element of the overall production outlay, far outweighed by savings in labour costs.
Investing in East Asia was no random decision. As well as plentiful, low cost and well-trained workforces, the region had good basic infrastructure. In 1945, for example, when Taiwan ceased to be a Japanese colony and was occupied by the forces of Chiang Kai-shek, it already had an extensive railway system, a widespread network of surfaced roads and a reliable electricity supply. The governments concerned were also for the most part strong advocates of economic planning which had the advantage for foreign investors of providing clarity and consistency in respect of policies. And, in contrast to many governments in developing countries at the time, notably in Latin America and India, for the most part they welcomed foreign direct investment.
Early projects were predominantly in Malaysia and Singapore, but three plants were set up in Taiwan, which encouraged the government to try to attract more. Keen too to see the local electronics industry become more valued, in 1973 the Taiwan government established the Industrial Technology Research Institute (ITRI) to pursue research and development in the sector, initially working closely with foreign investors. In 1980 Hsinchu Science Park opened, with a strong focus on advanced technology, above all servicing the requirements of the electronics industry. Deciding that the demand for, and importance of, semiconductors in the electronics industry was likely to grow significantly, the government designated semiconductor production as a strategic sector, with resources ploughed into research and development. The approach bore strong similarities to the way Silicon Valley in California had grown up around military research facilities and a Stanford University science park, and young Taiwanese software engineers were encouraged to spend time in Silicon Valley, often with government help, to learn about the latest developments and ideas. In 1980, the first dedicated Taiwanese semiconductor manufacturing company, United Microelectronics Corporation (UMC), was established by ITRI in Hsinchu.
Other Asian countries such as South Korea and Singapore were pursuing similar strategies, but Taiwan’s had one major difference. Crucially, the UMC business model did not envisage it competing with the still-dominant electronics giants. Instead, UMC would be but one of many suppliers to the giants of components for their final manufactures, albeit an important one. Buying in chips rather than producing them themselves enabled big firms to keep their prices low and margins high. Far from seeking to compete with these companies, UMC was happy to produce chips for them, to their specifications and designs. In 1985, UMC was spun out of ITRI and listed as a public company. Today it accounts for around 7 per cent of global chip production.
Around the same time, the government recruited a Chinese American, Morris Chang, from the US, where he had been responsible for Texas Instruments’ (TI) semiconductor business, to head ITRI. Two years later, Chang set up the Taiwan Semiconductor Manufacturing Corporation under ITRI auspices, using Taiwanese government funding and know-how from Philips. From TI, Chang already had a reputation for an aggressive pricing model in which chips would initially be produced at a loss while market share was built up. At TSMC he added to this an almost uniquely Taiwanese approach to business.
Typically, most business models, certainly in consumer goods, will seek from an early stage to develop brand recognition for the product in question with the aim of charging a premium price once the brand’s reputation is established. Taiwanese companies by contrast, with only rare exceptions, have instead been happy to act as contract manufacturers to others, making a product, or essential components for a product, that will bear another company’s name. The profit per unit produced is usually much smaller than at other firms, but Taiwanese producers offset this against their much larger output, and in this way can supply several companies with rival products, without competing with them.
This absence of brand names has meant that many Taiwanese companies remain all but unknown outside their own country. Prior to the pandemic, TSMC may have been able to claim to be the largest company in the world no-one had ever heard of, with a current market capitalisation of US$322bn, some way ahead of the better known Samsung of South Korea, and far ahead of Japan’s Sony. Of greater significance though, for both TSMC and Taiwan generally, is that as the complexity and capacity of semiconductors continued to expand so too did the cost of investing in their manufacture, especially of the most advanced versions. Today, the cost of a new fabrication plant or ‘fab’ for the most advanced chips is in the region of $20bn, a daunting sum for all but the largest or most specialised companies. The result has been that more and more companies were content to design their own chips but contract out the manufacture to TSMC, UMC, Powerchip Technology Corporation (another Taiwanese producer), or one of a small number of their competitors elsewhere.
Today, Taiwan accounts for 22 per cent of the global output of chips, the US just 12 per cent and Europe even less at 9 per cent. But these figures understate the real importance of Taiwan, for TSMC has a near absolute global monopoly on the supply of the most advanced chips. Other technology companies, including giants such as Nvidia and AMD, rely on it almost exclusively for their own needs, as does Apple for its iPhones and the US military for its most advanced fighter aircraft.
Such dependence is not without risk, for Taiwan especially, for whom semiconductors and related products now account for almost 30 per cent of all its exports. But there has been no magic about the industry’s rise to prominence. It can be summarised as demonstrating clarity of purpose, consistency in approach, and constant focus on detail. Food for thought for Western policymakers.