Govt's policy to build semiconductor ecosystem improves on earlier attempts

Can India blaze into the select club of countries that have semiconductor fab units? Last week, the Cabinet cleared an ambitious $10-billion incentive package to do just that. The government’s intention is to create a complete semiconductor ecosystem — fabs, home-grown chip design, outsourced semiconductor and testing (OSAT) facilities and compound semiconductor plants.     

If the fab plant takes off, India could join the list of Asian countries — Taiwan, South Korea, Japan and China — that currently control three-fourths of the world’s fab capacity (US is far behind with a 13 per cent share).  

The policy is ambitious. It offers a 50 per cent incentive on the project cost for Indian businesses to set up semiconductor fab and display fab plants — at least two each. It has also announced subsidies of up to 30 per cent on the project cost to support 15 OSAT and compound semiconductor plants (which make chips used in, for example, mobile chargers, electric vehicles and telecom networks). A design-linked incentive scheme has been announced to encourage domestic start-ups in product design.  

Clearly, the focus is on Atmanirbhar Bharat, built around the government’s intention to raise electronics production from $75 billion to $300 in six years. And this cannot happen until the country reduces its reliance on imports. Ashwini Vaishnaw, communications and IT minister, is candid: “Countries that do not know how to manufacture the semiconductor wafer, which is strategic, will not be able to progress.”     

This policy has been prompted by two key developments — first, the global chip shortage with little sign of let-up, and, second, the fear of dependence on China for fab supplies.   

But has this policy come too late? This is India’s third effort to set up fab plants. In 2007, Intel failed because the incentives and infrastructure were inadequate. In 2013, the government approved two mega semiconductor plants, supported by incentives, but companies (Jaypee group was one of them) did not have the money to invest.

Many players say India is late to the party when chip players are splurging on new capacity (including cutting-edge technologies) to reduce the chip shortage over the next 24 months (see table). The US alone has promised financial incentives worth $52 billion. “India should have gone full throttle two or three years earlier. It would do well to start talking and riding the next wave of investment,” said a senior executive of a global semiconductor company.

This year, global spending on setting up semiconductor plants will hit $146 billion, according to Gartner — 50 per cent more than last year — and these capacities will come up in the next few years. But more than half of the investments have gone to the US, which wants to increase its share of the global foundry base. The rest has been divided between Europe, West Asia and Africa (44 per cent) and Asia (29 per cent).

Vaishnaw said the government is offering global players something competitors cannot match and that is brain power. “We already have 20,000 design engineers working in global companies in India; 20 per cent of the global workforce in this industry is Indian. We have given companies a 20-year way forward, and we have supported it with start-up programmes for which we will train an 85,000-strong talent pool of engineers through universities,” he said.

Of course, India has the added attraction of large domestic demand, which is estimated to jump four-fold to $100 billion-plus by 2025. But Satya Gupta, advisor to the Indian Electronics and Semiconductor Association (IESA), said a plant requires $3-5 billion of investment and a break-even period of seven to 10 years, so no facility can depend only on domestic demand. It should be able to export at least 70 per cent of its capacity. “Also, domestic auto or mobile companies don’t buy chips directly from fab plants but from third parties such as Qualcomm or Mediatek. So the key is that these companies have to shift part of their third-party long-term production contracts, which are now, say, in Taiwan to India, and that takes a lot of time,” he pointed out.

This apart, few specialised equipment makers of semiconductor machines said they are inundated with pending orders and would take 18-24 months to make fresh deliveries. In other words, don’t expect a fab plant to hit India for two to three years even if investments are made today.

The upside is that time may give the government the opportunity to resolve one other key sore point for potential investors — the lack of ready-made infrastructure. In a recent interview, Intel India MD Prakash Mallya said a manufacturing ecosystem for semiconductors would require high-quality and uninterrupted power, logistics and water.

To be sure, the government has been proactive about engaging with top global and potential domestic investors. Talks have been held with the world’s leading chip makers — TSMC, Intel, Hynix, STMicroelectronics, to name a few — as well as with the Tata group, which has announced plans to set up a semiconductor facility. Vedanta had led the way, its Japanese company AvanStrate having tied up with Vedanta Resources and committed $15 billion in a display fab plant and semiconductors over 10 years, marking a revival of talks it had held five years ago for a $10- billion plant. South Korean giants are already investing $705 million to set up a mobile device display plant and talks have been going on to build a larger display fab plant, too.

Clearly there is a market for display fabs — demand is expected to double to $15 billion by 2025. And there are good reasons for setting up one in India from a strategic perspective. Seventy per cent of the LCD display market is controlled by China, which is also challenging the domination of LG and Samsung in the global OLED space. Given the geopolitical tensions, India can ill afford to depend on China for such a crucial and high-value component (which makes for a quarter of the cost of a mobile device and half for a TV).

Indian players have announced plans for OSAT and compound semiconductor facilities. Tata Electronics has hired Raja Manickam, an expert in OSAT business, who sold his company recently to Hero MotoCorp. The global market for this business is already $32 billion and is expected to grow six-fold by 2028 (the market is dominated by Taiwan). But India could leverage its cost advantage in the business.

The government is trying, for the first time, to encourage home-grown fabless product design companies to manufacture the chips they design in Indian foundries rather than in Taiwan, which could give them a cost advantage. Hemant Mallapur, co-founder of chip product design company Saankhya Labs, said: “The whole industry has only a few small players with total revenues of not more than $30 million. Designing a chip set is expensive and costs $ 2-10 million, and venture funds don’t fund us. So the government incentive will give us a huge boost to grow our business and move globally.”

The first big steps for India’s tryst with semiconductors have been taken. Whether it will be third time lucky is an open question.