The Hindu Business Line : To fab or not to fab...

Does India need a semiconductor fabrication industry or not? The debate is not new but the ‘yes’ and ‘no’ sides offered new insights into the issue at the recent ISA meet. The question is whether the government will get ‘bang for its buck.’

“We gave it up in Europe and it was wrong. There is no good reason why India should not have a fab industry.”

M. Ramesh

The debate splits the industry right down the middle. It’s been there for years and refuses to go away. It didn’t, at the recent conference of the Indian Semiconductor Association, where ‘which side are you on’ was a frequent conversation opener.

Does India need a semiconductor fabrication industry or not? By now, the basic arguments on either side are well known.

The ‘yes’ side points to the explosive growth in the Indian electronics industry — where chips are used. Some estimates put the size of the industry at $320 billion by 2020 and that’s a lot of silicon. Should India be dependent on an external source to produce the chips?

The ‘No’ side does not agree with this reasoning. In a global economy, the concept of self-reliance is dead, it says. Fabrication is an extremely capital-intensive industry — it takes about $3 billion to $5 billion to put up a fab, and, as technology advances, similar investments would be needed every five years to keep pace. India, an established player in design as well as ATMP ‘assembly, testing, marking and packaging’ should stick to the knitting. Fabrication can be got done elsewhere in the world. (see ‘Three sub-segments’).

This has been the essence of the arguments of the two sides. Admittedly, no new point has been brought to bear on the issue. However, at the ISA conference, protagonists on either side brought interesting insights, adding colour to their arguments.

A Gururaj

A. Gururaj, who is today Senior Vice-President, Manufacturing Projects, Reliance Communications, is a known face in the industry. In his previous job, he headed Flextronics, the EMS company that has put up a plant near Chennai. He says that while at Flextronics, his team did a careful study, going down to the minutest detail, and determined that a new fab in India would take not less than seven years to match its Taiwanese or Chinese competitor in price.

Dr Benno Fritzler

Dr Benno Fritzler, Vice-President, Base Technologies & Services, Infineon Technologies AG of Germany, does not subscribe to this view. He notes that Malaysia has been able ramp up fab units in quick time and they are not outpriced by the Taiwanese or Chinese. Dr Fritzler, a Ph.D in solid state physics, is all for taking the plunge. The fabrication industry, he says, “is not for the faint-hearted.” Yes, there are challenges, but you have to “start sometime.”

Malcom Penn

Malcom Penn, Chief Executive, Future Horizons, is in complete agreement. “Don’t give up,” he says. “We gave it up in Europe and it was wrong. There is no good reason why India should not have a fab industry.”

Rajeev Mehtani

Rajeev Mehtani, Vice-President and Managing Director, NXP Semiconductors, provides a good reason why India should not invest in fab — the need to make heavy bouts of investments regularly. Moore’s law says that transistor density doubles every 18 months or so — which basically means that either the chip will keep becoming smaller or will be capable of more and more functions. As such, a fab unit would need to keep investing to keep pace with the times. Very risky.

But wait a minute. No matter how technology advances, cannot a fab just produce the old generation chips? No, says Mehtani. He does not believe that there would be a market for old generation chips. Would you buy the kind of a bulky mobile phone that was produced in the 1990s?

Again, Dr Fritzler disagrees. He says there is always a market for the old generation chips, though perhaps for different applications. His company, Infineon walks the talk. When the world is moving towards semiconductors of 45 nm and below, Infineon has undertaken to buy 130 nm chips from Hindustan Semiconductor Manufacturing Corporation. “There is a big market for 130 nm chips,” he says.

Dr Brian Shieh

Nobody disagrees that fabrication is a capital-intensive business. Dr Brian Shieh, President, Powerchip Semiconductor Corp, Taiwan, knows this only too well. His company, one of the larger semiconductor companies in the world, reported a loss of $400 million last year. Yet, Powerchip continues to invest. That is the nature of the business. Not for the faint-hearted, remember? Dr Shieh points out that the industry is cyclical and stresses that you must have the guts to invest during a downtrend, so that you are ready for the upswing.

Shrewd readers such as yourselves may have noted one thing in the discussion thus far: all those ‘for’ are foreigners while all those ‘against’ are Indians. But that is not quite true.

Ashok Soota

For example, we have in a person no less than Ashok Soota, Chairman and Managing Director, MindTree Consulting, and former President of CII, a supporter of India developing a fab industry.

Craig Johnson

Don’t look at it in isolation, he says. You put up a fab, the whole ecosystem develops around it. And, Craig Johnson, Corporate Vice-President, Marketing & Strategy, Cadence Design Systems Inc, believes now is not quite the time for India to think of a domestic fab industry.

There are others who note that globally, the fabrication industry is losing money and except TSMC and UMP of Taiwan, nobody is making money. The industry is entering a phase of consolidation and India would do well to wait until the consolidation dust settles down.

Clearly, the stand on this issue is not dependent on the colour of the skin.

Jairam Ramesh

However, another pattern is evident. Those ‘against’ look at the business case, while those ‘for’ take a long-term, visionary view. The ‘visionary view’ was perhaps best articulated by Jairam Ramesh, Union Minister of State for Commerce. “There is a view that India does not need to invest in fab. We have rejected the view. To us, investments in fab are strategic investments.”

Government’s role

That brings us to the role of the government. The Government of India has promised to be a co-investor in the Indian semiconductor fabrication industry. In its semiconductor policy announced last year, it has promised to contribute between 20 and 25 per cent of the capital expenditure by way of equity, depending upon where the project is located.

Businessmen can break their heads over whether there is a business case or not. Investments and project viability are the industrialists’ headache and only business rationale can guide an investment decision. However, when public money gets involved, it brings the debate to another level: what are the tangible benefits to the people of India?

Again there are arguments on either side. Like this:

For: It is a strategic investment.

Against: Oh? What strategic? Don’t you import other essentials such as oil and fertilisers?

For: It helps build the entire ecosystem, which generates employment.

Against: Fab units do not generate employment. The electronics industry does. Electronic goods will be manufactured in India, fab or no fab.

For: A thriving economy should proudly have a domestic fab industry.

Against: It is like a teenager wanting to own a red Ferrai.

It goes on. There are people who wonder why the Government of India should willingly put in equity into an industry which will take a long time, if at all, to break even, when the funds could be utilised to support employment generating industries such as textiles, leather and small-scale units.

Dr Fritzler asks that if it is not a sensible idea, why would countries such as Malaysia or Vietnam give sops to attract the industry? And look who’s gone to Vietnam: Intel. Others are quick to point out that Dr Fritzler’s own company, Infineon, has taken a decision not to get into fab.

Some conjecture that the government knows that there may not be a need to put in equity funds because there will not be private investments in India in fab.

A fab unit is critically dependent on infrastructure — clean environment, plenty of water and supply of high-quality power — none of which obtains in India. So, is it a cold calculation that the policy would remain on paper?

That opens up another front for discussion. The semiconductor policy also covers a few other sectors, including photo voltaic wafers. Thus far, projects worth about $14 billion have come up on the drawing boards, which are entitled to benefits under the policy. But more than half of them are in the area of PV wafers. These include projects of Moser Baer, Titan Energy, Videocon, KSK Energy, Signet Solar, Surana Ventures and Terra Solar.

The social spin-off of government’s spends on these units is unclear. The contribution of these projects to solar power generation in India will be miniscule, as much of the production will be exported. Even if it is not, a couple of thousand megawatts in a country that has an installed power capacity of 1,60,000 MW and wishes to add 80,000 MW in the next five years is but a drop in the ocean. Government putting equity money in these projects? What is the point of it?

ramesh@thehindu.co.in

Illustration: Satheesh Vellinezhi

More Stories on : Hardware | Insight | Events

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
To fab or not to fab...