At Microsoft's research center in a leafy lane in India's tech capital, a new generation of researchers are being groomed half a world away from the software giant's sprawling headquarters in Seattle.
Complete with beanbags and coffee served in steel tumblers, the center is helping change the perception that India is no place for top-end research and development.
Staffed with about 60 full-time researchers, many of them Indians with PhDs from top universities in the United States, the center is at the cutting edge of Microsoft's R&D. It covers seven areas of research including mobility and cryptography.
Its success, including developing a popular tool for Microsoft's new search engine Bing, underscores the potential of R&D in India at a time when cost-conscious firms are keen to offshore to save money by using talented researchers abroad.
Showing off the Bing tool which enables searches for locations with incomplete or even incorrect addresses, B. Ashok, a director of a research unit at the center, said the innovation would never have taken root if the R&D had been done in the United States.
It was completely inspired by the Indian environment, but is applicable worldwide, he said.
While India might seem like a natural location to expand offshoring into R&D, it is hampered by some serious structural problems that range from not enough home grown researchers to a lack of government support.
India produces about 300,000 computer science graduates a year. Yet it produces only about 100 computer science PhDs, a small fraction of the 1,500-2,000 that get awarded in the United States, or China, every year.
Students here are not exposed to research from an early age, faculties are not exposed to research and there's no career path for innovation because there's a lot of pressure to get a 'real' job, said Vidya Natampally, head of strategy at the Microsoft India Research Center.
With few government incentives and an education system that emphasizes rote learning, India lacks the kind of environment found in say, Silicon Valley, where universities, venture capitalists and startups encourage innovation.
China has a policy in place for R&D; we don't, Natampally said, adding that India could move up the value chain faster if even a small percentage of its engineering graduates went into research.
The small numbers of PhDs and the lack of government incentives for India's fledgling R&D sector are blunting the country's edge, analysts warn.
Rival China has already pulled ahead with more than 1,100 R&D centers compared to less than 800 in India, despite lingering concerns about rule of law and intellectual property rights.
Aside from providing funding to encourage students to complete their PhDs, China also offers fiscal incentives such as tax breaks for R&D centers and special economic zones provide infrastructure for hi-tech and R&D industries.
India is also losing out in the patent stakes. In 2006-2007, just 7,000 patents were granted in this country of 1.1 billion people, compared to nearly 160,000 in the United States.
We're nowhere near the U.S. or even Israel when it comes to innovations, said Praveen Bhadada at consultancy Zinnov, which estimates the R&D sector in India is worth about $9.2 billion.
Our costs are low and our talent pool is ahead of China, Russia and Ukraine, but China gives specific incentives, and produces way more PhDs than we do.
India is cheaper than China for R&D, those in the industry in Bangalore said. But salaries in India have been rising by about 15 percent every year and may soon reach parity with China. R&D center costs in Shanghai are currently just 10-15 percent higher than in India.
Microsoft and other firms have been working around the government's indifference.
Cisco, IBM, Intel, Nokia, Ericsson and Suzuki Motor have all gone beyond low-end coding and tweaking products for the local market, with hefty investments and recruitment.
Their success shows India's potential if the government starts supporting such ventures and building high-tech parks and incubators.
If Paris asks for some work, it's not because they think it's cheaper but because they want inputs from India, said Jean Philippe, chief designer of the Renault India Studio, which competes with the French carmakers' five other global studios.
Texas Instruments and San Jose-based Cadence Design were among the first to set up R&D in India in the mid-80s, drawn by the legions of English-speaking software engineers who could be hired at about 20 percent of the cost of engineers in the United States.
The opening of the economy in the early 90's and the establishment of the software services industry drew more foreign firms looking to cut costs and tap emerging markets.
From when a few companies offshored non-critical design work, we have seen India emerge as a preferred destination for design and development of chip, board and embedded software, said Jaswinder Ahuja, managing director of Cadence India.
Firms first focused on the 'D' in R&D, but research has grown in importance in recent years, and many of the facilities in India are now the largest outside their home base.
Half of Cisco's core R&D work, including innovations in WiMAX and optical networks, and about 40 percent of SAP's ideas for processes and product development come from India.
The Indian units are more tuned to the needs of customers in emerging markets. Besides, Bangalore is only a 5-hour flight away from three strategic regions: Southeast Asia, east Asia and the Middle East, said Aravind Sitaraman, vice president at Cisco.
IBM's India Research Labs do a fair share of patenting, helping swell the parent's record numbers every year, said director Guruduth Banavar in Bangalore.
Its new $100 million-mobile communications research, Mobile Web, is the first time a big project has been driven from outside the United States, he said.
For a research lab it's the best environment to be in: you can see the problems and the opportunities, said Banavar, who was previously at IBM's lab in Boston and has, like several of his peers, returned to India to oversee operations here.