The Hindu Business Line : `Now's the time'

IT is difficult for one person to be seen as mild and aggressive. Rajendra Khare manages to be both. He is a picture of politeness and courtesy when you interact with him on a personal level. But get him talking about the opportunity for India to grab a fair share of the semiconductor manufacturing pie, and you'd be surprised at the passion that comes through. Khare spent time with eWorld after the ISA summit. Excerpts from the chat:

Why is semiconductor manufacturing so critical to India?

India has a good value proposition. Right now, we are in design services, using intellectual capability to make the rest of the world successful. We are being paid wages, which would go down once others compete on cost arbitrage. There is hence a compelling proposition for semiconductor manufacturing. Then, the value of software industry would further grow on value, not cost arbitrage.

Margins are to be had in fabless units and design services. In manufacturing, margins are low, companies have long gestation period. We have advantages in software. Are we spreading ourselves too thin instead of concentrating on areas of benefit?

That is a view of the small picture. Look at the big picture. If you want to become an engineer, you don't look at one test and say what is my return on investment into this test alone.

If there is a ripple effect on the economy, then commitment from government not on building on infrastructure but actual cash investments in companies on hand, seems to be important. For, out of 10 fabs in China, we understand that only two are making money, that too only recently. Even for a player like Intel, the government has not yet given clear commitment.

We are not taking a one or two year picture at all. In this industry, which is mature around the world, there is some investment required from the government. Let us say that government sets aside about $2 billion for investments either in equity or to provide debt funding, over the next X number of years. The ripple effect is that you will see FDI worth $30-50 billion coming in because you are enabling infrastructure. If the government invests in that manner, you create a domestic industry that contributes to 12 per cent of GDP. In the country, you invest in national highways. We expect returns only in 10-15 years. This is like a semiconductor highway. Government has to put in money because it has high returns. Semiconductors are the oil of the information age. You cannot look at just the next 1-2 years and ask "Where does the government have that kind of money... ?"

In China - whichever sector - it's been mandated by government. We want private sector to take over in every sphere, without the government committing some resource...

If you can encourage local consumption, that would make fabs profitable. Fabs are like aircraft and hotels. At 90 per cent occupancy, you are profitable. At 30 per cent, you aren't. If you can ensure capacity, then you are profitable. How do you ensure capacity? Within the country itself, you consume $15 billion worth semiconductors. The market would dictate what would be produced in large volumes here. When you consume such large volumes, and when product selection is right, filling that capacity won't be so difficult.

You talk of a narrow window of opportunity. Why?

Several reasons. One, because of momentum built up, investors are sitting out there.

These are private investors with shorter time lines, a 3-5 year window. Wouldn't you consider partners such as TSMC and UMC, who would be better committed...

It's a race. You can take one option as the best and waste all your time trying to convince them. Or you take a value proposition to one who is hungry for doing this... first come first served. Now, if you lose the momentum, the next time around, investors will be suspicious of false starts. Next time they become hot, it would be 5-8 years. The last time the steam was built up, it was as early as 1990. Korea and Taiwan had no fabs at the time. Now, the momentum has been built so you don't want to diffuse it.

India has a unique opportunity now. There is new capacity being built up around the world. Technology transition is happening from 90 to 65 nm. The next opportunity would come when the next transition happens, when new capacities would be built up. Money for that comes from around the world. A lot of investments go to China due to lack of alternatives. The whole world cannot look at only one option. And new capacity is going to be built. If India does not stand up, someone else will. Investors are not looking only at India. But, India has other advantages: strengths in design and software - it becomes triple play with manufacturing. We have the intellect. You cannot create that overnight for other countries. Whereas manufacturing is a matter of investment. You can create it in three years. Over time, if you can build on the design strength, then we would be the number one choice. It's a leadership position we are looking at 10 years from now.

There is tech obsolescence in manufacturing. So why not wait for the next cycle?

No, no, no. That won’t work. It takes 10 years to put up something. At the time, we would be in the same position. Then, of course the cost would have inflated. State-of-the-art plant now costs $3 billion, in three years, the same would cost between $3-5 billion. We would do the same song and dance about why not wait for more time. Now is the opportunity. We are latecomers who have to be in a tearing hurry. If you keep waiting, the distance keeps increasing.

Assuming SemIndia alone invests and others don’t come in to create that cluster, do you really think that it would be viable?

Now, here’s why a cluster would come. Like SemIndia, others are also interested in talking about fabs. In the next 3-5 years window, if the government provides this incentive, and along with this we have a strong market, with demand being strong, people will come. For, government’s incentive won’t be available after 3-5 years.

We had the same opportunity 10 years ago with contract manufacturing, except the Tandon Group, no one else really picked up, despite all the relaxation that would have been done.

If you do things before time, you don’t succeed. Timing is important. Market has respect for you only now. Earlier, you were hardly consuming. Now, you are a big market, with advantages in design and software strength. Earlier, manufacturing was independent of those two. Now, integration is happening.

Cost of entry is $2 billion – that would create opportunities that would make this look a very inexpensive investment.

If Intel decides to invest in India that would be significant. Any others?

We don’t have to worry about who is investing. What we have to worry about is that new capacity should be created here. It is a $235-billion industry today. This industry has to manufacture chips somewhere in the world. It always needs to add new capacity every three years. That capacity should be somewhere. We could be the destination for that, whoever invests. People should get excited and come here.

The $2 billion you have asked for, set aside, is there a certain maximum the government would invest in each company – say, $200 million in 10 companies, etc…Up to 26 per cent stake is the maximum. If you put up a $3-billion plant, equity is only one-third. The investment would maximum be $250 million. Look at the leverage. By putting in $250 million, you have a $3-billion worth of economic activity going on in the country. Some fabs may come in with older technology. Their investment would be only $500 million. So you can create that cluster.

If you take on too much debt, would it not be excessively leveraged?

Issues we have faced earlier – 1994 – markets were so in any sector. If anything goes wrong, it has a cascading effect on the whole economy.

It does have a down side too. The semiconductor industry is cyclical. But if you time it when it is on the rise, you make your money. The industry has had 15 per cent average growth in the last 10-12 years.

Is the government then looking at short term? It has apparently refused even the $100 million that Intel was rumoured to be asking for?

I cannot comment about Intel. But players need to bring in technology, not merely assembly plants. Also, policy cannot favour one player and not another. The intention of the government is that high-tech fabs should come in the country. And it cannot be blackmailed into doing this or that. We are not beggars. We are providing an opportunity. Interested people should come and grab that opportunity. We should not go out begging with a bowl. And the government realises the opportunity we are sitting on.

The F&S report – how realistic is it? Is it far too optimistic?

They have done their job in an extremely professional manner. A study of this kind in the industry in India has not been done before. The scientific precision with which they have gone about it is rare to find. They have found the demand side – people who consume – got a figure. Then they went to the suppliers to match those figures. They have gone to the extent of dividing the industry into 52 application segments. Then they took the bill of materials for these. Any study could go wrong. But here, they have fallen on the conservative side, rather than be optimistic. Most studies are done with little data. This has so much data collected and analysed between June 05 and January 06. Initially we thought we could do it in four months. Sometimes market agencies do a lot of extrapolation but here it was an industry body involved. And we really wanted accurate data. F&S rose to the occasion.

Fab city blueprint. Is it waiting till the budget end?

We have already made the blueprint and that is what is in discussion with the government. They need to see the full picture. We engaged with consulting partners who are tax experts, technical information, etc. They need to know what this animal is. The blueprint tells you what path to take.

More Stories on : Interview | Events | Hardware

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

Stories in this Section
It makes sense to share

The die is cast

Drawing on software

When one complements the other...

Power the process

Too snoopy for comfort, say some

Backup for e-mails in Outlook Express

Spike buster

`Now's the time'

Driven by convergence

Quiz

The art of being a CTO

Cartoon