The Hindu Business Line : India inside?

Indian vendors now have a chance to deliver complete products to semiconductor users, signalling the convergence of hardware and software companies and their growing interdependence in the semicondcutor space. Or do they? eWorld moderates a debate.

Do Indian software companies targeting the semiconductor industry now have better opportunities to do high value work?

Take, for example, mobile phones. Several Indian software companies each develop a part of the software that makes a phone work across the world. But what if one company had the expertise and the marketing strength to make and sell these software designs whole, instead of piecemeal? And what if we had several such companies targeting different areas, across telecom, consumer electronics, computing, power industry and the like? Sounds implausible? Or obvious?

To debate this, eWorld, along with the Indian Semiconductor Association (ISA), got together a panel of experts comprising S. Janakiraman, Chairman, ISA, President and CEO, R&D Services, Mindtree Consulting Ltd; Prof Sadagopan, Director, International Institute of Information Technology - Bangalore (IIIT-B); Raghu Panicker, Sales Director - Mentor Graphics, India; Surendra Jain, Managing Director - Sequoia Capital India; Sanjeev Keskar, Country Manager, Sales, Freescale Semiconductors; Dr S. Karthik, Engineering Director, Analog Devices India; and Poornima Shenoy, President, India Semiconductor Association (ISA). K. Bharat Kumar moderated the discussion in the run-up to the annual ISA Summit conducted in February each year.

Excerpts:

Mindtree offers a wireless USB software product in its entirety for mobile phones. Could you take us through the experience of making this in India and selling it to the world?

S. Janakiraman

Janakiraman: With the economy opening up in the 1990s, Indian companies, which were already selling hardware in India, took a look at the global markets too. That is how this entire service has become the design service industry. In addition, we had the multinationals setting up their captive centres and getting designs out for the global market. All these have led to proliferation in India along the design services value chain. In the semiconductor area, we all do work on architecture verification or typical design. Product companies do work here - be it in board design or embedded software or protocol engineering. The subsystem level of engineering service is the focus.

What I see happening these last three or four years is we are required to add much more value rather than just thrive on cost arbitrage. Today the clients want us to help them reach out to the market better. Here, three things are happening. One, Indian companies have strongly aligned with semiconductor companies as alliance partners;

Take Freescale or Analog Devices as examples. They have good semiconductor chips and look to take services to the market - be it automotive market or a video surveillance system. Design services companies are able to make reference designs and integrate software, are able to showcase their technology in market and in that process we are able to sell more semiconductor components. Hence we sell more services.

Two, today customers do not want to reinvent the wheel. They want to reuse components, intellectual property or a software building block which is already fabricated and can be plugged in. For a mobile phone company the next focus is on how to make the next great phone. Blue Tooth is an essential element. So they source that kind of technology from a Wipro or a Mindtree. Three, is the `fabless ODM model'. Earlier, companies used to outsource only manufacturing. Now, they outsource the entire engineering to us. It's not just the electronics or just the software; it includes mechanical, compliances, the complete product.

Analog Devices is doing some interesting work in India. Your views on the emergence of product design opportunities?

S. Karthik

Dr Karthik: We are a captive design arm for our parent. We are into IC design, product test engineering application support to customers. In the last five years we have expanded our work. The entire product chain is done here. We work with thirdparty vendors, too. Semiconductor products are done here; that is one shift that I have seen. All of it, from product definition, design, implementation, what happens when the chip comes back, silicon debug, validation, customer support. the entire thing is done here. For that you would need a lot of technical depth. Also, this is a long-term process. We need engineers who have burnt their fingers with many chips. They must be able to resolve customer issues. It is not a one-time affair. If you have 10 years as the chip's lifetime, you need to support this for at least 10 years.

Your company targets the Indian market with its products. How are the acceptance levels?

Dr Karthik: One example of reasonable success is our offerings for the energy meter market. We had chips particularly for the Indian market. Specifically for India, we have a tamper-proof feature. This is not available in other countries. For this, we need to work with partner companies and provide them a lot of support so that they make a fine product before it is delivered to the end-user. Partners would make the box and integrate the software. The Electricity Board would source the box from the electronics companies.

Did demand come to you or.?

Dr Karthik: Typically there is a demand; you have a base product and then you figure out that there are market requirements. Then, you evolve that base product to suit those requirements. You always have some technology code and you say `ok, this could be applied to this.' and so on. The big change I see is that in the 1990s, we used to take to global markets what we did in India. Now, we are customising global offerings for the Indian market.

Freescale sells to, among others, the automotive industry in India. What are the opportunities here?

Sanjeev Keskar

Keskar: Domestic demand is fuelling design activity in India, especially across three verticals. Indian companies developing products for the Indian market. Design services companies are making products for global customers in India. Companies such as Honeywell, Dell, GE, Bosch and Visteon are developing products for India as well as for export.

According to the ISA-Frost & Sullivan reports published in August 2007, the total consumption for semi-conductor which was $2.7 billion in 2006 is going to be $5.4 billion in 2009; we are going to witness a growth (CAGR) of 26.7 per cent over three years.

Globally, the semiconductor market is growing in single digits. We have opportunities in five verticals:

IT and office automation - nearly 6 million desktops were manufactured in India last year;

Telecom - we have seen BSNL expand and Nokia and Motorola setting up plants in Chennai;

Consumer gadgets - the large consumption of TVs and audio systems, set-top boxes boxes is evidence;

Industry - where we have seen 10 million energy meters consumed by electricity boards and a large consumption of UPS inverters and motor controllers. (The tech evolution is evident in the inverter. Three years ago it was just a power electronic gadget; two years back there was an 8-bit application in it; today we see DSP controlled inverters);

Automotives - seven million two-wheelers and 1.4 million four-wheelers are made in India; Three years ago, Bajaj used to give one microcontroller for one vehicle; but now they are giving three - one for electronic batch board or cluster; one for body electronics and the third for ignition.

The electronic design analysis (EDA) companies supply tools to chip designers. You also provide simulation tools. An engineer can demonstrate a capability overnight instead of waiting for three months to run a prototype. Your perspective?

Raghu Panicker

Panicker: First, companies in India sourced chips. Then came product development and now we actually conceptualise the product. We are doing everything; India cannot marginalise itself to be just in services.

EDA companies add value by helping with verification. The faster the verification, the quicker the product reaches the market. How important is it to have companies across the chain located in India?

Dr Karthik: Now we have the presence of all EDA companies in India. There are large chunks of development in India, in addition to support and services. I can now access the designers, and their support R&D is also here. Product developers as well as their customers are in India. They can quickly flush out products for India by talking to the design community here. That's the evolution in the last 10 years. I am happy that my designers need not have to spend much time for tool support. We are closely connected with the EDA companies. If they want to expand into product development they have us and they can flush out new product ideas.

How has the model evolved this last decade?

Janakiraman: Ten years ago, to get a new electronic design was a nightmare; you spent millions of dollars even without knowing whether you will succeed or not; in the last 4-5 years, software tool companies are talking about the subscription model where you don't own the tool but you subscribe for a month or a quarter or a year; the next stage was where toolmakers got their fees as a percentage of revenues of the client, instead of paying for the tool. The whole business model is going through a healthy change.

Poornima Shenoy

Poornima: EDA companies are physically located where the customers are. They are in substantial strength across Bangalore, Hyderabad and Noida.

Janakiraman: It looks trivial now, but earlier, to get a tool fixed in a machine it was a real nightmare. You had to call Santa Clara, wait for two days and sometimes you won't get the attention. Today it is next-door; you sometimes get the EDA engineer working in our teams so that they can improve the tool.

Is there hesitation from venture capitalists to invest seriously in Indian design firms? Is there a gestation period needed before we see VC activity here?

Surendra Jain

Jain: I think you have to consider this from a larger perspective by taking other examples and not restrict it just to the semiconductor industry. I would say that we have to evaluate based on the people who have some good, interesting ideas and then see if the ecosystem is willing to look at the risk-reward equation. There are three issues here:

Talent - there are enough people at different levels; it is not a matter of quantity but a question of quality. Now we see a talent pool and there are many experienced people becoming available. It is only a question of time;

Market - the ecosystem developing here is important. But the market itself needs to change. What do people like you and me buy? It is beginning to change;

Capital - India is clearly an attractive investment destination. Capital is being made available, especially during the last seven years. I think with that we are sure to succeed. In turn, this would encourage others to look upon successful ones as role models. Innovation is important. The Tata Nano is a great example. Take Caf‚ Coffee Day. Coffee and tea were available everywhere in the country for Rs 2. But here it's not just coffee but a whole experience. That is innovation. On the semiconductor side, it is more a question of time.

India churns out 30 PhDs in Computer Science. The figures for the US and China are 1,400 and 3,000. What should the education industry be doing to make it easier for the software companies targeting thesemiconductor market?

Prof Sadagopan

Prof Sadagopan: Until a few years ago, if you finished your PhD in India you could look forward to academic positions only and not industry position. My students tell me that they like me a lot but that they do not want to teach. To add to that, a master's graduate enrolling in a PhD programme gets Rs 7,000 per month while his counterpart in the industry gets more than Rs 25,000. It was a socioeconomic problem. I think it is changing now. People complete PhDs and are in the industry as academics, with companies with R&D centres such as IBM, Motorola, Yahoo, Google and HP. So, three years ago we in IIIT-B started this: We asked Infosys to give the candidate a scholarship of Rs 25,000 per month. We got three such scholarships. Now we have 14 such - these are high-value credentials. This would lead to PhD numbers growing.

What number is ideal for this industry?

Prof Sadagopan: Around 100 PhDs per year, and probably we will see this moving to the above-500 levels over the next few years. When PhDs work with the likes of Google and Microsoft and are happy with the money they get, they tend to pass on the message. Success breeds success.

Is there enough risk-taking ability among our entrepreneurs for this kind of work?

Janakiraman: Risk/reward is an element of, one, mindset and, two, the ecosystem. The latter has changed now - today there is money available and there are people willing to invest in risk-reward models; they are also aware of the changes that have happened globally in the Silicon Valley. So, why not over here?

Next is the mindset - whether you want to take the risk or not. I think it is not suffi- ciently there yet. People want to take lowrisk, medium-reward kind of option. The services industry traditionally has been in that category only; there are many services that have come up in India. But there aren't too many product companies. There are exceptions like Tejas Systems, Ittiam Systems and their ilk. I am eagerly awaiting the day when Tejas would have an initial public offering.

We need to create awareness at the academic levels. As a nation, we are very bookish and understand a lot of things as theory. But we do not make risk-taking ability as a virtue.

How is the demand scenario looking?

The world's leading mobile phone companies are those that could well do with outsourced engineering. But we see only piecemeal work being done for them in India.

Is it tough to get a foot in the door?

Janakiraman: The rhythms in the products and design services are entirely different. Companies such as Wipro, Mindtree and HCL should experiment with the product as an independent entity rather than under the design services umbrella. In design services, 95 per cent of the people are technical. For, the essential ingredient is the knowledge and people.

In a product company, normally 15 to 20 per cent is invested in R&D. The remaining goes to manufacturing, channel, brand and support.

If you are a good design services company, it doesn't automatically mean that you can be a good product company.

Today we build prototype technologies and then license the product to the customer, rather than wait for the customer to tell us what to do.

So, we create that kind of VC funding from within. Companies could collaborate on this front. It is evolving now. In five years' time it should be prevalent.

Thanks to B Visweswaran, Central Marketing Group, Sasken Communication and Arun Natarajan, Venture Intelligence, for special inputs to this article.

More Stories on : Interview | Hardware | Software

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

Stories in this Section
Icing takes the cake

No dumb machine

Dive deep for expertise

Windows boot error

Quiz