Overcoming increasing semiconductor entry barrier

| Updated: July 27, 2022 22:04:26

-Reuters file photo -Reuters file photo

The increasing volume and role of semiconductor, as silicon chips, have been alluring growing number of firms and nations to enter the semiconductor value chain. Like iron and plastic, silicon chips are essential in making all kinds of industrial products. Silicon chips have been migrating from peripheral roles to the core functioning of products, which were not considered high-tech even just a decade ago. Furthermore, silicon chips-powered cumulative effect of incremental innovations has been resulting in creative destruction, leading to migration of innovation epicentre-causing a transformational impact on industrialisation. But unfortunately, although the journey of semiconductors over the last 70 years has been showing no sign of slowing down, the entry barrier is exponentially growing.

Six hundred billion dollar semiconductor value chain has a few major segments such as (i) production of highly pure silicon, (ii) electronic design tools for designing chips, (iii) designing chips, (iv) photolithography for projecting designs on the silicon wafer, (v) processing for creating 3D nanostructure of the design, and (vi) testing, bonding, and packaging. Starting from logic circuits to memory, the semiconductor industry produces an array of different components. Even the primary technologies used to make silicon chips are being used to produce displays and core component of LED light bulbs.

Once we talk about the high volume of semiconductor business and envision its growing role in shaping economic activities, aspiring entrants tempt to jump into the idea of setting up a chip fabrication plant, known as fabs. Unfortunately, both the complexity and investment needs have been growing. For example, despite having money, access to technology, and a market, Intel has been struggling for more than a year to migrate from 10nm to 7nm node. On the other hand, the investment need for setting up a fab has been exponentially growing, reaching $15 billion for a 5nm node.

Hence, among other countries, India has opted to invite foreign direct investment (FDI) to enter the semiconductor value chain. Unlike in the past, tax differentials and preferential access to the domestic market, expected to grow as high as $110 billion by 2030, are not good enough to attract such FDIs. Hence, staggering cash subsidies are being to allure companies like Taiwan's TSMC or America's Intel. According to Nikkei Asia, India is ready to fork out $30bn for alluring Taiwanese firms and other foreign players to realise its ambition of developing a domestic supply chain of semiconductors. From this vast pot, around $10 billion are for two chip facilities and two display plants. Government of India is reportedly willing to provide support of up to 50 per cent of the project cost to eligible companies.

On the other hand, to take a pie of the growing 3D NAND flash memory market, which is expected to reach $78 billion by 2030, China has channeled $24 billion to build a domestic firm. Furthermore, countries with a solid semiconductor base are also pumping staggering public funds to retain the position. For example, the USA has come up with a $52 billion package to get back the silicon edge.

To make it worse for aspiring less developed countries, the scope of value addition out of labour has been shrinking-reaching less than 10 per cent. As opposed to a few simple workstations fitted with optical microscope and ultrasound bonding machines, once labour-intensive outsourced semiconductor assembly and testing (OSAT) are demanding more capital investment and less labour. For example, Tata has been after setting up such a plant for $300 million. But through a humble way, Intel started such a facility in the paddy field of Penang in 1971. Barriers of entry in semiconductor equipment and fabless segment have also been exponentially growing. For example, DUTCH ASML entered the lithography segment through modest investment with 31 people in a few wooden huts. But the same company invested more than $10 billion over 23 years in developing a 13.5nm machine to monopolise the high-end photolithography market. On the other hand, Japanese Nichia had to make a Nobel Prize-winning scientific discovery to create the LED chip business.

As explained, the entry barrier has been growing despite the growing consumption of semiconductor components. Some of the significant factors are (i) growing investment need, (ii) increasing economies of scale and scope, (iii) rising science and technology content, (iv) decreasing labour content, and (v) rapid obsolesce. But does it mean that countries those are now in a strong position faced a similar situation to make an entry? Ironically, NO. They have been taking advantage of these factors for increasing the competitive edge, monopolising every segment of the value chain and making the entry barrier increasingly higher. Furthermore, a few of them made entry with as little as $25,000-that is the money Sony paid to Bell labs in 1952 for licensing fee of Transistor to kick start Japan's semiconductor industry. On the other hand, Dutch national Arthur Hendrik del Prado paid virtually nothing to give birth to the semiconductor industry in Europe, which has culminated in the global monopoly of ASML in lithography, making it the 42nd most expensive global company.  

Taiwan's glittering success story, with more than 65 per cent share of the global contract chip-making market, also began humbly. It started with a $4 million contract given to American RCA to set up a training and research fab in 1974. On the other hand, without any government handout, Intel happily started semiconductor bonding and testing plants in the paddy field of Malaysia's Penang. Over the last 50 years, both these humble beginnings have grown into large-scale success. But unlike Taiwan's focus on knowledge and ideas, Malaysia kept focusing on labour. As a result, once poorer, Taiwan has become far more prosperous than Malaysia.

With the given unfolding situation, what are the options for aspiring less developed countries? Should they get into the race of offering increasing cash incentives and protection? Highly likely, smaller countries will fail to keep pace with big ones. Instead of jumping into building infrastructure and getting into the race of incentives, they should pay attention to understanding and monitoring the unfolding dynamics. To make an entry and create a growing profitable business, aspiring countries should focus on developing a solid intelligence base for detecting opportunities and getting ready with adequate capacity to be at the right place at the right time. For example, as opposed to getting into head-on competition with American or Japanese firms or alluring them to invest in Taiwan with incentives, Taiwan focused on intelligence-leading to entry into 3rd party pure fab market, which was ignored by American, Japanese, and European firms. Furthermore, making for import substitution is no longer a wise option due to growing economies of scale and scope, and decreasing labour content.  Hence, the focus should be on detecting discontinuity and building the capacity to add value through the fusion of labour, knowledge, and ideas in creating a snowball effect-as opposed to winning the race of giving incentives and protection.

M. Rokonuzzaman, Ph.D is academic and researcher on technology, innovation and policy.

[email protected]

Share if you like