New chips on the block7 min read . Updated: 18 Jun 2010, 10:52 AM IST
New chips on the block
New chips on the block
As the world grows fatter and richer, cardiovascular diseases will kill more than one third of all people by 2030, predicts 37-year-old Balasubramanian L.
A beleaguered, hypercompetitive global automotive industry will not tolerate a failure rate of more than two parts per million, explains Prem Kumar Arora, 28.
Mukul Ojha, 31, discusses the lighting options urgently needed for 65 million Indian households that light homes with kerosene.
It’s a balmy, breezy Bangalore day outside. Here in the air-conditioned meeting room—higher than the tops of the old rain trees on the edge of a technology park—it’s hard to think of common ground for such diverse data.
Try semiconductor devices, computer chips laced with silicon: nerve centres of the electronic devices that run our world, from steel factories to cellphones, from cars to healthcare instruments.
Also Read More articles in the series
All the bright, young people in this room have degrees from the Indian Institutes of Technology or the Indian Institutes of Management, or both. Vice-presidents may be in their early 40s, directors in their late 30s, and in this rare Indian tech story, many in this building run global divisions for their company, 28-year-old Cypress Semiconductor Corp.
As demanding customers squeeze profits, demands for new technologies spiral, and markets move east, Cypress is moving critical functions from California to India.
Bangalore is now home to not just half of Cypress’ global chip-design resources but also marketing, global business development and the company’s largest test facilities. Global strategy and mergers and acquisitions are run out of Chennai.
“We’ve got a lot of control and ownership," says Sumeet Shantanu, who, at 33, heads global business development for the memory business out of India. “I can solve any problem in India, not go to San Jose."
Arora, senior product marketing engineer for automotive semiconductors, explains how his team works on 50 projects across the US, Germany and Japan, nimbly tailoring products to customer needs. For 12 years they sold memory devices that controlled unsexy stuff like fuel-air mixtures.
Today, Arora smiles as a control panel destined for a car’s navigation map or air-conditioning control reacts to a wave of his hand. “The entire automotive human-machine interaction strategy was conceived in India," says Arora. The idea evolved in 2004, production began in 2006, but Cypress expects first returns only by 2011.
Cheaper, faster, smaller
The semiconductor industry lives, and dies, by a simple motto: cheaper, faster, smaller.
“Thanks in large part to fierce competition and to new technologies that lower the cost of production per chip, within a matter of months, the price of a new chip can fall by 50%," notes a semiconductor industry analysis by Forbes Online.
Falling prices can actually mean growing costs for semiconductor companies.
Apart from customer pressure pushing up research and development expenses, it costs $5 billion (Rs23,250 crore) to set up a chip-making factory or fabrication unit, a “fab" in industry parlance.
So, cameras, cars and cellphones may change every six months, but for semiconductor companies, the first dollar on a new product can take up to three years. “If you want to be profitable today, you need a 50% gross margin (to survive)," says Cypress’ senior vice-president, India operations, Rajeev Mehtani. That’s 50 profit cents on a dollar.
“Right now, we’re coming out of the industry’s worst recession since 2001," he says. The downturn sent the industry’s annual growth rate from 12-15% to 3%, so there’s very little room for error.
Ten years ago, a cellphone had 273 components. Today, it has 99 or less.
All this is possible because of a law first postulated in 1965 by Intel Corp. co-founder Gordon Moore, who said transistors—building blocks of modern electronic devices, used to route and amplify signals—squeezed onto a chip would double every 18 months. The average cellphone today has more computing power than Apollo 13, the US spaceship that first put man on the moon in 1969.
The leading edge today is 40 nanometres. That’s an electronic circuit less than one-thousandth the width of a human hair, packed with one million transistors.
Thanks to the explosion of computing power, semiconductors are, at heart, a commodities business, like pepper or iron ore. Unlike commodities, getting the chips to market isn’t the challenge. It’s what you do with them.
Of the world’s 200-odd semiconductor companies, only two now have “fabs", Intel and Samsung. The others have moved to “fabless manufacture", or outsourced chip production. Most companies such as Cypress use transistor-packed, disc-shaped silicon wafers from third-party “fabs" as raw material to create integrated circuits.
These brawny circuits power the global economy—and pose a new challenge.
“The rate at which we’re putting transistors on a chip is much faster than designers filling those spaces," explains Mehtani. “We’re wondering what the hell to do with all the real estate on a chip."
At 46, Mehtani is a 23-year veteran of the profession and very much the elder statesman in his office. He seeks to push Cypress into a host of new applications and markets, using a philosophy called more-than-Moore—instead of simply cramming more transistors on a chip, evolve clever uses for chips by lacing them together and telling them what to do.
At Cypress, one-fourth of all engineers are software experts, figuring out smart, new uses for semiconductors.
India may have rebounded from the global downturn, but in the semiconductor world, things are grim. So, Cypress uses the energy of these young Indians, throwing them off the deep end into a thrilling ride where they get to steer.
“The challenge is what we get here," says Surya Koneru, 40, analogue design engineering director, who drives Cypress’ 60% global market share in touch-phone technology. Half of Cypress’ 300 engineers are now in Bangalore, its largest technology centre, travelling the world in a relentless search for customers.
Putting these young men and women in charge has led to the industry’s lowest attrition rate: 5%.
“Nowhere else do we get this kind of exposure," says Koneru. “Nowhere else."
“We did it in this building"
A floor below the meeting room, it’s four degrees cooler than the building’s 24 degree Celsius. Over furrowed brow and broad grin, Cypress’ engineers and technicians tackle technological demands and those wafer-thin margins as they ponder innovations for new—and as yet tiny and unprofitable—markets.
“This is where the rubber hits the road," says Karthikeyan Madathil, 38, business development director, recalling a Chinese customer who wanted an image sensor for a foodgrain-sorting machine. Since prices of foodgrains such as rice and peanuts are based on colour, shape and size, the machine’s electronic sensor needed to be “very, very, very fast".
Last year, says Madathil with pride, “people in this building completed the job".
It’s a phrase you commonly hear. While they are part of a multinational operation, Cypress’ young team takes distinct pride in their achievements as Indians.
Silicon absorbs light in pretty much the same way as a human eye. That allows Cypress’ engineers to create electronic eyes for a variety of uses, from satellite sensors and security systems to the latest movie cameras.
“Some of the best digital movie cameras in the world have sensors designed right here," says Madathil, “in this building".
There’s also evidence of gambling in this building—on new products for emerging markets.
Senior product marketing manager Balasubramanian heads a Cypress project for one such device for India: a single-chip machine that can measure temperature, blood pressure, blood oxygen and heart rate.
With its burgeoning prosperity and an increasingly sedentary lifestyle, India is ground zero of a global surge in cardiovascular diseases and diabetes. An all-in-one apparatus could have a huge market.
“It is definitely doable," says Balasubramanian, who follows a plan—increasingly adopted by semiconductor multinationals in India—of seeding technology and courses in semiconductors into colleges and later reaping the expertise.
So far Cypress has set up 19 laboratories in Indian colleges; semiconductor companies have set up more than 530 labs across the country, mostly in the south. Cypress’ target for this year is 30 new labs and 50 research projects.
About 10km east of Cypress India, Aravind Sitaraman,47, vice-president and managing director at Cisco, a $36 billion US tech multinational, also believes India operations must change gears.
“There is nothing like working with a customer and co-creating a product, then replicating it across the world," he says, and sees energy, education and healthcare as the next opportunities. “In IT we have an excellent business model, but is that the endgame? I don’t think so. We have to develop unique products across the world, take risks. That is coming in, very slowly."
Cypress’ faith in local talent does not guarantee success, even if it follows a progression that benefits India.
What helps companies like Cypress even in a downturn is the ability to hedge bets as electronics spreads deeper into daily life.
“With employment of 0.5% or less even among mature developed economies, the (semiconductor) sector’s direct contribution to GDP (gross domestic product) is limited," says a March 2010 study from the McKinsey Global Institute. “But ongoing innovations in the sector have contributed to the IT adoption that has improved business process and boosted productivity in many other sectors—and therefore made a difference for economy-wide growth."
Started operations (in India): 1995
Made in India: iPod wheel, human-car interface; most complex logic chip with 100 million transistors; 60 patents