Long-term success of fabs hinges on government policy

Asia’s success in semiconductor manufacturing is best illustrated in the numbers—11 of the world’s 14 top plants that produce close to 80% of microchips sold in the world are located in the region.

add_main_imageBold entrepreneurship and innovation are universally cited as the reasons for Asia’s success in emerging as a hub for semiconductor plants, or wafer fabrication units.

But the history of the fabs, as semiconductor plants are known, in Asia reveals the key role played by the state, from Taiwan to China to South Korea and Singapore. Put simply, the state led the creation and incubation of the semiconductor industry across the region.NextMAds

India’s semiconductor policy rightly aims to emulate the steps undertaken by governments in its neighbourhood in funding the risks, providing incentives and subsidies so that companies can replicate the success.

Although about two decades late, India’s attempts to enter this segment coincides with the shutting down of the last couple of fab facilities in the US and Europe and increased investments in this sector in Asia. It also comes at a time when chip making is losing its allure in many parts of Asia as young engineers prefer careers in software.

While Asian governments have adopted different strategies, the development history of the Taiwanese semiconductor industry provides the best lesson for India. Taiwan, which accounts for about 70% of the global “foundry” business, is also proof that a government-led foundation, be it a strong industrial policy, or planning and research and development (R&D) funding or even setting up pilot agencies, can translate into success for entrepreneurs.

At the same time, Taiwan also received the rub of the green. In the 1970s, as the Japanese made deep inroads and began successfully challenging the US monopoly in the semiconductor industry, leading firms there such as Fairchild Semiconductor International Inc., Texas Instruments Inc. and General Instrument Corp., among others, began moving their packaging and testing operations to Taiwan to cut costs. Apart from cheap labour, these companies also took advantage of Taiwan’s liberal foreign direct investment regime as well as its policy of duty-free flow of raw materials.

But when the attempts of independent Taiwanese firms to establish a semiconductor industry failed, the government got into the act, and in 1973 set up the Centre for Electronics Industry Development and Research (later changed to Electronics Research and Service Organization—ERSO) and tasked it with promoting the development of the industry. In the years that followed, Taiwan made the strategic decision of investing heavily in the semiconductor industry.

Taiwan’s semiconductor industry took off in 1975 when government research institutes set up the first homegrown manufacturer—United Microelectronics Corp., which soon became the world’s second largest wafer foundry.sixthMAds

This plant also created a pipeline of technical talent and entrepreneurs who went on to build new companies. In 1981, state-owned research and development units helped establish Taiwan Mask Corp. and over the next couple of years, ERSO spun off a slew of companies, including Taiwan Semiconductor Manufacturing Corp., which has grown into the world’s largest contract chip-maker today.

Interestingly, the Taiwanese government did not get involved in the manufacturing or in the running of these plants, but transferred the technology to private enterprises and limited its role to providing R&D.

With tens of firms being spun off with R&D support from the state, US and Chinese investors joined local businessmen in setting up fab units, even as leading global players in the sector began forming joint ventures with Taiwan’s upcoming manufacturers. This propelled the country into a new era—from imitator to global leader—in this space by the turn of the last century.

Taiwan’s emergence in this industry also led to semiconductor firms across the US, Europe and Japan outsource their manufacturing and concentrate on designing semiconductor chips. By 2000, from Intel Corp. to Applied Materials Inc., all leading chip companies in the world either had manufacturing facilities in Taiwan, or had outsourced the making of their chips to fab factories located in the island.

What should not be overlooked is that in addition to R&D facilities, the Taiwanese government’s contribution also extended to friendly investment policies, building infrastructure—electricity, water and other public faculties, and parallely creating a robust education system to tap the career opportunities provided by the newly set up fab facilities.

Taiwan’s success also led to its companies investing in China, which helped boost the latter’s efforts to become a semiconductor manufacturing hub.

In sharp contrast, the Korean model saw the government supporting the semiconductor industry only after the private sector, especially Samsung Electronics Co. Ltd, demonstrated it could break the Japanese and American dominance. But from the late 1970s, all South Korean companies concentrated solely on Dynamic Random Access Memory (DRAM) chips and currently have 70% of this market.

It was the state support to the Korean conglomerates—the chaebol—that led these companies, Samsung, LG Electronics Inc. and others, to diversify into electronics and later into semiconductors. The late 1970s saw South Korea’s leading chaebol investing heavily in electronics, during which time its government’s focus was on sectors such as chemicals and steel.

The chaebol initially expanded into the semiconductor space as their core products, be it televisions, electronics, cars and computers, among others, needed these chips. Samsung’s success at making a 64K DRAM in 1983 and 256K DRAM in 1986 provided the much needed boost to the industry and led to several Korean conglomerates entering this segment. Its government then stepped in with subsidized funding for R&D and product development.

Just like Taiwan, Korea, too, benefited from the trade wars between Japan and the US in semiconductors in the 1980s. The major break came post the 1986 Semiconductor Trade Agreement that placed curbs on Japanese DRAM exports to the US, allowing Korean companies to get a foothold in this space.

The trade agreement also forced Japanese companies to develop next-generation DRAMs, forcing many of the firms to completely quit making lower-end models and ceding this space to Korean manufacturers. With active government support, Korean manufacturers led by Samsung played the price game in the 1990s as the diversified business of chaebol allowed them to cross-subsidize their semiconductor offerings, gaining further market share.

But when the Japanese responded with similar price cuts to kill off the Korean firms, the impact was felt by American DRAM makers, who could not compete at these prices and went bankrupt. The chaebol then bought up several of the American firms that were folding up and enhanced their R&D and design capabilities to emerge as the undisputed leaders in this segment, a position that they still hold.

Yet, not all state-promoted models have been successful in Asia. For over a decade, in the 1990s and the early 2000s, the Chinese government invested billions in setting up state-led chip-making faculties, including buying costly production lines from abroad and training employees outside the country with the hope of becoming Asia’s semiconductor powerhouse.

But it was only when entrepreneurs, many of them who had come back to China after successful careers abroad, entered the scene post-2000 that China’s semiconductor industry began to come to life. China today accounts for about 10% of the total production of semiconductor products and is looking to increase its global share to 15% by 2015. A close analysis shows the state may be the biggest barrier to Chinese firms’ failure to play a global role in the semiconductor space, leading to a trade deficit of nearly $150 billion annually in this segment. This is because, unlike Korea and Taiwan, state-led efforts in China continue to be aimed at creating indigenous technologies and protecting its small semiconductor units rather than global collaboration and integration, which is vital for investment and technology transfers.

Even if India’s ambition to manufacture chips takes off, its long-term success will depend on government policies that will provide incentives to leading global firms to move R&D and design facilities to the country. Consider the case of Singapore, a country with one-third of the population of either Delhi or Mumbai. The city state has about 40 design centres, among them, nine of the top 10 chip design companies.