July 09, 2023

Massive Subsidies to MNCs Not a Road to Self-Reliance in Semiconductors

Bappa Sinha

MODI Government has made a lot of noise about making India self-sufficient in semiconductor production. The Micron deal was announced with much fanfare last week as "historic" and proof of PM Modi's vision finally bearing fruit. A closer look at the deal shows India footing 70 per cent of the 2.75 billion dollar deal while getting Micron to set up a low-end chip assembling and testing facility, while Micron gets 100 per cent ownership of the facility.

Last year the government announced a 10 billion dollar scheme for encouraging manufacturing as part of the "Make in India" initiative. Three parties had expressed interest in setting up semiconductor fabrication plants in response to this scheme. The most hyped of these was the Foxconn/Vedanta signing an MOU with the Gujarat government for setting up a 20 billion dollar chip fabrication plant. Those grand announcements have now come to nought with the government rejecting the promised 10 billion dollar incentives for this project as neither Foxconn nor Vedanta have the required chip-making experience and are yet to find a technology partner who has the necessary technology.

Are these massive corporate subsidies to foreign multinationals justified, and would they lead to self-reliance in semiconductors and computer chip-making? Let us take a closer look.

Electronics and the semiconductor chips which power them are critical to modern industry and economy. They are essential to everything: from consumer electronics, control systems for automobiles, industrial machinery, aircraft, computers, smartphones, satellites and military hardware to emerging areas such as IOTs, AI, 5G and robotics. The global electronics industry is estimated to have a revenue of around 3 trillion dollars/year. Within that, the semiconductor industry had a revenue of around 600 billion dollars last year and is estimated to grow to around 1 trillion dollars by the year 2030.

Given their importance to the economy and the military, the electronics and semiconductor industry is of strategic national importance. We have seen in recent years how the United States is seeking to impede China's economic progress by trying to restrict its access to high-end semiconductor chips and the technology required to manufacture them locally. Therefore, governments in many major economies around the world have recognised this importance and are formulating policies to strengthen their indigenous semiconductor industry and also to ensure their continued supply.

India has almost no semiconductor and chip fabrication capabilities at present. This is the result of the mistaken belief which had taken hold during the post-liberalisation era of the 90s that hardware was easily available in the global markets, and India should focus only on software and systems as a national strategy. This was different from its earlier policy of self-reliance, in which the indiginisation of the entire supply chain was planned. It was then recognised that semiconductors and electronics is a strategic sector, and without this, our space, atomic energy and defence industries could be seriously weakened.

India attempted to establish self-reliance in semiconductors starting in the late 70s and early 1980s. The Indian government identified microprocessors and other semiconductors as potential foundational technologies for the future. In 1984, the Indian government funded a 100 per cent State-owned enterprise – the Semiconductor Complex Limited (SCL) in Mohali, Punjab. SCL's goal was to eventually design and manufacture leading-edge circuits and electronics and provide the foundation of India's electronics industry. The government invested around 50 million dollars into the venture, a significant amount of money at that time.

SCL hired top engineering talent from IITs, IISc Bangalore and other institutes. They also managed to recruit from leading Indian electronics companies like Bharat Electronics Limited, the Indian State-owned aerospace and defence electronics company. 

SCL went about acquiring older technology, realising that there is no point in starting from scratch and reinventing the wheel. The idea was to strike a technology transfer deal and then get trained in the new technology, hoping that with good human capital and sufficient financial backing, it could improve the technology and catch up with the market leaders: known as a fast "seconds" strategy. 

At SCL's founding, it was able to license a 5-micron process technology from American Microsystems Inc. AMI was not a market leader like Intel or Motorola, but it was a start. SCL was able to acquire process technology from two other companies – Rockwell and Hitachi. These technology transfer agreements allowed SCL personnel to get trained in the new technology. They augmented this trade knowledge with academic partnerships.

SCL advanced very quickly from the 5-micron process technology down to a 0.8-micron or 800 nanometers process in the late 1980s. This was just two years or one technology cycle behind the market leaders such as Intel. It seemed possible that India could achieve its goal of being a global semiconductor manufacturer within a decade. Then disaster struck.

A major fire broke out in SCL, and the entire plant was burnt down. Some suspect arson and sabotage as the cause of this fire. This struck a devastating blow to India's self-reliance hopes in semiconductors. A blow from which we never recovered. As India entered the liberalisation era, subsequent governments never seriously attempted to revive the project, even though lip service was paid to it by multiple governments.

SCL in the early 1980s had the advantage in that everyone else in the rest of the world was not too far ahead of them. Taiwan and China had yet to move into semiconductors. And the equipment back then was a lot cheaper. Today's leading-edge chip factories regularly run up to more than ten billion dollars. And all that investment has to pay for itself in a highly competitive and volatile electronics world.

The chip-making process is also highly complex. It depends on cutting-edge advances in science and technology. It involves multiple steps in its supply chain, all of which need to work together to develop and produce the final output, the chips. These can be classified into:

1.    "Fabless" chip design is done by companies to develop circuitry for new chips and appliances.

2.    Electronics Design Automation Software (EDA) tools are used to design, analyse and verify the design of chips and circuits.

3.    Manufacturing equipment used for chip production, which includes very high-end lithography machines for etching circuits on Silicon wafers. These machines are the Extreme (EUV) and Deep Ultraviolet Lithography (DUV) machines from ASML, Lam Research's plasma etching machines and Applied Materials' eBeam metrology system for semiconductor patterning control.

4.    Fabs or Foundries where the chips are actually manufactured. Almost all the highest-end chips, those below 10nm, are produced by only two companies in the world: TSMC (Taiwan) and Samsung (Korea).

All these diverse building blocks of the semiconductor chip supply chain are developed and produced in a handful of advanced economies, including the US, EU, Taiwan, Korea, Japan and now China. No single country controls the entire supply chain.

China launched an ambitious "Made in China 2025" industrial plan – introduced in 2015 and designed to reduce dependency on foreign technology – the Chinese government has committed enormous sums to subsidise the construction of fabrication facilities and advanced manufacturing capacity, setting an objective of expanding semiconductor production to meet 70 per cent of domestic demand by 2025. US and EU have also announced significant stimulus packages for their semiconductor industry in order to try to arrest and recover some of their lost ground in semiconductor manufacturing. Simultaneously, the US has also announced sanctions against China, seeking to restrict its access to high-end semiconductor chips and semiconductor manufacturing technology as a means to hinder China's economic progress. It is in this background of intense global competition that the Indian government has also announced its plans to encourage foreign companies to set up semiconductor fabrication plants in India.

The semiconductor space requires very high capital investments as the manufacturing equipment is very expensive. As an example, a single Extreme Ultraviolet Lithography machine from ASML which is required to make the highest end 7nm and below chips, costs 125 million dollars. To further complicate matters, semiconductor technology has a very high rate of obsolesce, with each generation of chips becoming obsolete in five years or less. This means that plants set up to produce chips need to be upgraded almost every five years at very high capex costs in order to stay on par with other players. All these challenges make it very hard for new entrants to enter the space. 

Along with high capex costs, this space also requires very high R&D costs as the industry depends on the latest advances in science and technology. As a result, for a country to advance in this space requires significant investment in research laboratories in the university system and research institutes. China has significantly invested here and has outpaced even the US in producing both Science, Technology and Mathematics (STEM ) graduates, and patent applications. India, on the other hand, has seen an intensification of neoliberal reforms in the higher education sector with significant cuts or stagnation in funding for universities and research institutes.

Last year, the government announced a 10 billion dollar scheme for semiconductor and display manufacturing as part of the "Make in India" initiative. The government sought to attract at least two semiconductor fabs and two display fabs. They have emphasised 45nm technology, thereby indicating that they are looking for a reasonably newer technology. The 45nm is still four technology cycles – or around ten years – behind the global leaders today, but it could be considered a reasonable starting point today. The government will provide 50 per cent cost for fabs and 50 per cent capex support for other steps such as design and ATMP (assembly, test, marking, packaging) as an outright subsidy.

It remains to be seen if such generous subsidies will spur foreign companies to move their semiconductor manufacturing to India. However, even if they do, it will not mean that advanced technology will get transferred to us automatically, and it will enable us to develop self-reliance in this field. Without insisting on technology transfer agreements and simultaneously investing heavily in higher education and research institutes, self-reliance will not automatically happen. This requires planning and coordination with various research institutions and the industry for a path towards eventually achieving self-reliance and an Indian semiconductor industry. Instead, with the government's current vision of "Make in India", the emphasis is on inviting foreign companies to set up factories in India by enticing them with overly generous subsidies, cheap labour and extremely lax labour regulations. This is not a path toward an independent semiconductor industry in India but a path towards continued dependence, that too for technology that would be years behind the US, China, Japan, South Korea and Europeans.