|
Tessolve founders Raja Manickam and V Veerappan rush into their new, regal
office building at Bangalore's Electronic city, as a Dataquest team awaits a
facility tour. “We were held up at a meeting with Wipro,” Veerappan tells
them apologetically. In less than 24 months of opening up shop, the first post
silicon testing shop in India, his life has been quite eventful, busy as he is
developing test programs, cut opening chips, analyzing failures, speaking and
'chipping' at seminars. In his lab stand blue-coated men and six
machines-testers to be precise-each one of them costing about $2 mn, or Rs 9
crore. “In India, we don't touch anything that is capital intensive,” he
blurts out. “This is a capital intensive business.”
Tessolve's resolve sounds like the trumpet of changing times. At the other
end is Vinod Agarwal, who now heads a company initially founded in the US by
entrepreneurs and semiconductor experts that took up the challenge of putting
together a consortium. His SemIndia fab, to be located in Hyderabad, will be the
country's first major attempt at semiconductor manufacturing. Projected
investment: $3 bn.
 In a way, both Vinod and Veerappan complete an unfinished agenda, an
ecosystem that in semiconductor parlance is called the 'food chain'. India
had design and software. In between the two, sits manufacturing and testing. The
food chain can now possibly feed the electronics industry better, which, in
turn, will increasingly cater to one of the world's biggest middle class, a
very young population with rising per capita income and the world's fourth
largest economy on a purchasing power parity basis, behind the US, China, and
Japan.
The optimism about India's place as a major semiconductor hub, a one-stop
chip shop, grew manifold after the Indian semiconductor Association (ISA) and
Frost & Sullivan held up to light, buttressed, and re-buttressed some
well-known and a few not so well-known facts in an exhaustive study: India's
rapidly expanding GDP over the next several decades will boost electronics
demand-the total electronics production as a percentage of GDP has been
increasing from 1.5% in 2000-01 to 1.7% in 2004-05-the total electronics
production in 2004-05 was aggregated at $11 bn, up from $9.7 bn in 2003-04.
Expected to reach $58bn in 2010 and $155 bn in 2015.
|
|
| At his VLSI chip testing
lab in Bangalore, V Veerappan,
co-founder and director, Tessolve, and his team of 50 engineers, probe the
wafer after it comes out of fabs from across the world-Tessolve also
does production testing |
The shared optimism is because in certain electronic products, such as
advanced cell phones, for instance, the silicon content could be as high as 60%.
Such phones will have a chip for its camera, the MP3, power management, and
memory among others. The semicon content could also be as little as 5%. Taking
an average 20% silicon content, the semiconductor industry's domestic revenues
should swell to a moderate $31bn in the next nine years -the study estimates
2005 revenues to be about $ 3bn, projects it to grow to $12bn by 2010 and $ 36bn
by 2015, at a CAGR of 29.8%.
| In the last
24 months, practically every chip company has set up a design center in
India. Out of the top 25 chip companies, 18 are already here, among 125
design operations in India |
Place in the Sun
India, as the next great semiconductor market opportunity or as the next
dream destination, is, therefore, no longer a fragment of anybody's
imagination. But for the more patriotic, completion of the 'food chain' is
also strategic. The semiconductor world is going through the next paradigm
change, it is said. The initial paradigm was that of Integrated Device
Manufacturers (IDMs) such as Intel, which did everything under one
roof-designed, manufactured, wrote software, and sold it. The fabless model
made its entry next and design became separate from manufacturing. Independent
foundries started to manufacture. Fabless companies such as Broadcom started to
do a lot of design. Lot of IP got subsequently developed, chips that earlier
used to be a small block in a system, now became the system. The product,
because it was a system, contained huge amount of software. The embedded
software became the third axis in this three dimensional game.
| Semiconductor
Design: Identity Crisis |
|
Where does design
fit in when you're charting industry segments and revenues?
A lot of the work that happens at Intel (IIDC) or Texas Instruments is
simply termed “software and services”. Even though the work can and
does also involve hardware design, most of such work is classified under
“services”, though there are cases where hardware design may be
included under manufacturing contracts, for instance, and be captured
under hardware exports, etc.
However, not all of
these 'services' may be getting captured under the software and
services revenues, as estimated by Dataquest as well as Nasscom.
The ISA-Frost and
Sullivan research on the Indian semiconductor sector pegged the design
market to be about $3.5 bn. This includes VLSI, hardware/board design and
embedded software space. Many semiconductor companies in India are members
of NASSCOM and therefore have been submitting their design revenues to the
body. A large part of the embedded software revenues, which ISA quotes at
$2.5 bn, thereby does reflect in NASSCOM or Dataquest Top 20 figures. But
not all of it, and the discrepancy could be as much as a billion dollars. |
|
|
|
“If we have an
electronics hub in Chennai, chip manufacturing in Hyderabad, and design in
Bangalore, we have a wonderful opportunity to become the design capital of
the world”
-Poornima Shenoy, ISA
President |
But design and manufacturing, that became independent earlier, is now
becoming inter-dependent again. “That means you cannot design a product unless
you know enough about manufacturing. And you cannot manufacture unless you are
working proactively with design,” explains Raj Khare, MD, Broadcom India. A
system on chip today, has more than 10-12 thousand programming registers.
Writing software for this kind of a system is impossible unless one works up
ahead with design. And if one doesn't know the end application, he cannot
design.
This is what he calls the triple play paradigm. India is good in software. In
the last four-five years, India's ability in semiconductor design has gone up
multifold with companies even taking complete responsibility for an entire
product. What we don't have today is manufacturing. “So if we don't invest
into manufacturing right now, our ability to design will go down because it is
becoming interdependent,” Raj warns.
Manufacturing, therefore, becomes strategic. A look at the world-wide
scenario explains why. The US, which used to be a complete player, is no longer
competitive in manufacturing-wherever electronics manufacturing happens,
semicon manufacturing follows-and in the US, electronics manufacturing will
not be cost competitive. Large-scale migration of design has already happened to
India. Some Apac countries are strong in manufacturing, but not so much on
software and design. In the next 10 years, India can thereby logically become a
complete triple-play paradigm player. It is difficult for Apac overnight to
create an intellectual capital. Taiwan doesn't have scalability for
intellectual capital because it is a small country whereas manufacturing is
something investment-based. Japan is a very high cost manufacturing destination
and has an uncertain future. India, on the other hand, has a cost-competitive
structure, comparable to China-a country that has the language and IP barrier
to climb. A lot of fab investments are presently going to China because of the
huge infrastructure they have created. But the point is, the world always needs
new manufacturing capacity to be built. If India positions itself as a
manufacturing destination, a good chunk of investment will come its way. The
time is now, and, as Raj says, “They cannot put all eggs in one basket.” Page(s) 1 2 3 4
|
Print
Comment
Email
Digg
Del.icio.us
Reddit
Twitter
