 "IndiaAI CEO Abhishek Singh")
India must develop capabilities across the entire artificial intelligence (AI) value chain to create a self-reliant AI ecosystem, rather than remain a consumer of global AI services, Abhishek Singh, chief executive officer of the IndiaAI Mission, said.
He made the remarks on the third day of the TiE Global Summit and the regional AI Impact Summit here on Tuesday.
“In the entire AI stack — whether it is energy, data centres, compute, models, or the applications layer — capabilities must be built at all levels. Currently, India has strong expertise in building applications and use cases, but we need to invest across the entire stack to ensure that we become AI service providers, rather than just consumers of AI services,” Singh said.
To address these gaps, the government is supporting all layers through the IndiaAI Mission. India currently has 38,000 graphics processing units available, of which around 25,000 are operational and being used by startups, industries, and academia, he added.
The datasets platform AIKosh offers nearly 6,000 datasets for unrestricted use, Singh said, adding that any Indian startup can access these resources, which include contributions from government departments, public institutions, and private companies.
At the same event, Vishal Dhupar, managing director for India and South Asia at Nvidia, said that while the country needs to build its own large language and AI models, it should also leverage its unique characteristics to innovate for efficiency and inclusion.
“There is a narrative in India that we don’t build frontier models, and that we will be the capital of use cases — and that’s fine. It is an alternative to the Chinese and American approaches, which I would call the Indian way. This approach reflects India’s diversity: officially 22 languages and 1,500 dialects, a massive population, and very different affordability levels. When transactions are large, but values are small, innovation is essential,” Dhupar said.
On training young people for emerging roles in data annotation, analysis, and data science, Singh said around 570 data labs are being established nationwide at institutions such as industrial training institutes and polytechnics.
“We need to train people in data science, data annotation, coding, and understanding how AI works. Many of these data labs are located in smaller cities,” Singh said.
Dhupar also praised India’s talent pool, saying the country is home to some of the finest engineers and programmers.
