What if you could hear the Internet, instead of reading it? Search by talking, instead of typing? Researchers are working on hyper linking speech to create a 'Spoken Web'. It could change how the world communicates. And it's happening in New Delhi, not Silicon Valley. Fortune India brings you a rare peek into India's top labs.
ON A 50-ACRE CAMPUS IN Whitefield, Bangalore, researchers work on projects ranging from mechanical engineering to polymer science. The sylvan setting-trees, lawns, a Japanese garden-is protected by a 12-foot-high wall, electric fences, and security cameras, giving it the feel of a top-secret army installation. Some areas, where Indian researchers work on classified projects of the U.S. government, have even more security. Welcome to GE's largest multidisciplinary R&D facility-its first outside the U.S.
The state-of-the-art centre is an example of how R&D has evolved in the country. There was a time when MNCs turned to India for cost arbitrage, tinkering to adapt high-tech products for local use, and taking on the tedious bits of overseas projects.
No longer. The country is an increasingly important R&D hub for global projects. Fortune India spoke with around 40 researchers and industry analysts, and there's no debate: Labs in the country are creating new technology that has worldwide impact. Intel's energy-efficient Atom microprocessor was developed here, all Motorola networking equipment uses Indian-made software, and Texas Instruments India has played a key role in developing the LoCosto chip for cheap mobile phones.
According to Bangalore-based management consultancy Zinnov, which focuses on R&D, the number of multinational companies R&D centers in India rose from 191 in 2000 to 780 in 2008. Most were set up by U.S. companies and focus on information technology. Barring the Netherlands-based Philips and Japan's Suzuki Motor, R&D investment from other countries is low-key.
Up until the mid-1990s, most research in India took place in some 400 government-funded labs and focussed on national concerns such as nuclear physics, space research, public health, and food security. Industry participation, mostly by MNCs, has been a catalyst, linking research to product development and giving it the momentum lacking in government-funded research. India plans to increase R&D spending to 2% of GDP by 2012, up from just under 1% today. (China spends 1.5% of GDP, Japan 2.4%, and the U.S., 2.8%.) That means more than doubling R&D outlays in two years to approximately $23 billion (Rs 1.061akh crore), assuming 8% GDP growth.
According to the Department of Science and Technology, between 2001-02 and 2007-08, private-sector spending on R&D almost quadrupled, while state spending doubled. Over the same period, R&D expenditure by the private sector as a percentage of total spends moved up from 19% to 30%. With a few honorable exceptions, such as Dr Reddy's Laboratories, Indian companies don't invest much in R&D. So it's safe to assume that it's MNC labs that are doing most of the spending.
Zinnov estimates that 5% of MNC labs in India lead global projects. IBM's Spoken Web project, for example, is led out of Delhi. On Hewlett Packard's Bangalore campus, researchers are working on technology that recognises human gestures and eliminates the need for a keyboard. DuPont's and GE's India centres are working to develop nanocells that capture solar energy. Besides, there are partnerships such as DuPont's collaboration with Reliance Industries on polyester technologies and third-party R&D such as HCL's work on Boeing's 787 Dreamliner aircraft. "Indian campuses are no longer out-posts; they're mature enough to handle advanced projects," says Sudhakar Ramakrishna, corporate vice president, Motorola Inc. The company's Bangalore R&D centre, set up in 1991, is working on its 4G network management software, which it has developed from scratch.
When U.S. tech companies first set up R&D centres in India-starting with Texas Instruments' Bangalore centre in 1985-researchers lacked product development experience. Companies got around the problem by sending over Indians who were working for them in the U.S., mostly in Silicon Valley. In 1997, Adobe Systems sent Naresh Gupta, a Ph.D. from the University of Mary-land and an expert on web publishing and e-learning tools, from San Jose to Noida, to set up an R&D centre there. The following year, Srini Koppolu returned from Redmond to set up Microsoft India Development Centre (MSIDC) in Hyderabad. Then, in 1999, Intel sent Texas educated computer engineer Praveen Vishakantaiah to start Intel India, Bangalore. They were the vanguard of the talent that returned home years after leaving India to study and work abroad, and have shaped their companies' R&D activities in India.
Cost arbitrage today is more relevant at lower skill levels than higher ones. For example, a leading U.S. firm with an R&D centre in India says that at entry level, engineers here cost about Rs 8lakh ($17,000) a year-30% to 40% of their U.S. counterparts. Bangalore-based G.N. Udiaver, associate director at Manpower India, a recruitment firm, says the salaries become comparable as experience increases.
Even when the primary motive for companies to come to India was cost arbitrage, the vision was clear: to develop leadership locally. "We took it slowly," says Koppolu, who played a key role in developing Office 95 and Office 97 in Redmond, and recently quit after 21 years with Microsoft. "We started with debugging software, but were clear that we'd eventually do complete product engineering here."
If the reverse flight of talent let companies scale up quickly, being located in a fast-growing and diverse economy spurred them to the next level. IBM picked India for the Spoken Web-which will run on handsets, rather than PCs or laptops-not only because of the quality of its techies, but also the size of its telecom market. Many companies use their labs here to develop products for other emerging markets. "It's arrogant to think about products for those markets sitting in a developed country," says Wido Menhardt, chief executive officer, Philips Innovation Campus, Bangalore. "One has to be here."
MSIDC's emerging markets lab is a good example of this. "We try to make computers more relevant to customers in emerging markets," says Srinivasa V. Thirumalai Anandanpillai, the lab's group programme manager. One way is to facilitate use in languages other than English. His team focusses on solutions such as onscreen keyboards, transliteration, predictive input and mobile interface. It's already possible to use a qwerty keyboard to type in 10 Indian languages, and MSIDC is working to extend the solution to more.
Many companies, including IBM, Microsoft, and Yahoo, separate pure research from product development to some extent. Basic research identifies a broad problem and creates technology for a solution, which a product development team then makes relevant to the market. For instance, Microsoft Research India, Bangalore, found that conventional mapping software did not work well in India, as addresses don't follow regular patterns. It developed algorithms that an MSIDC team in Hyderabad then used to improve Bing Maps.
Another sign of R&D maturing in India is the growing room for creativity. Take Google, which set up R&D centres in Bangalore in 2004 and Hyderabad in 2005, to capture India's growing Internet user base. Today, 70% of its engineers' time goes into improving core products and services such as search, advertising, and apps, and 20% into strategic products such as Android and the Chrome OS. But engineers are encouraged to use 10% of their time to generate ideas. If an idea gets enough votes-as Google Finance did out of Bangalore-it turns into a project and eventually a product.
Recruitment at MNC labs has been rising. GE's Bangalore centre today employs 4,300 researchers and engineers who work on just about every product and technology of the company: It started out with 275 in 2000. According to Zinnov, all labs combined employ close to 173,000 re-searchers today, up from 145,000 two years ago.
Patents filed in India have more than tripled from 10,592 in 2001-02 to 36,812 in 2008-09. Their success rate (patents granted as a percentage of those applied for) has improved as well, from 11% to 44%. These numbers include patents filed by government labs and universities, but not those filed abroad for work done here.
It's not just about developing commercial products, but also about scientific endeavour. Private labs are increasingly collaborating with academic and government organisations, says P. Anandan, managing director, Microsoft Research India. Rather than being secretive about its research, his company encourages its scientists to publish their work in academic journals, and to be an active part of the global scientific community.
Kiran Karnik, former president, Nasscom (National Association of Software and Services Companies), says India should encourage R&D through incentives such as higher tax breaks and better infrastructure. "Knowledge is our differentiator, but we lack a framework to facilitate private investment." He argues that more might have been achieved had the state given higher priority to R&D.
While Karnik's concern is valid, for the moment nobody is complaining. As Bangalore-based Sridhar Sarathy, vice president at Juniper Networks India, says: "Some competencies at the Indian centre are equivalent to those at the headquarters. We're exploring what more can be done."
IBM BETTING TWICE AS BIG ON INDIA
TWO OF WHAT IBM CALLS "BIG BETS' (exploratory high-investment research programmes) are led by its India team. They are the mobile web, which includes the Spoken Web, and service quality improvement. The latter focusses on the needs of increasingly huge and complex service systems. As part of this Big Bet, IBM India has already developed a technology that analyses voice, so that a call centre executive can gauge a customer's mood even before taking the call. Another example is data cleansing, which analyses records and removes errors that lead to poor customer service, such as billing discrepancies or a wrong address that may delay a shipment.
IBM Research - India, set up in Delhi in 1998, innovates based on market insights, such as the challenges posed by irregular addresses. Its India Software Labs (ISL), established in Bangalore in the same year, takes care of product development. "The basic technology research happens in Delhi, but this would be of no use unless it is turned into products for the market," says P. Gopal Krishnan, vice president, ISL.
-- Fortune India, September 2010