Davos: India clearly belongs to the first group of AI nations as AI architecture comprises five layers — application, model, chip, infrastructure, and energy — and we are actively working across all five, Union Minister for Electronics and IT, Ashwini Vaishnaw, has said.
Outlining India’s approach to Artificial Intelligence (AI) the minister emphasised large-scale AI diffusion, economic viability, and techno-legal governance, during a panel discussion at the World Economic Forum (WEF) annual meeting in Davos.
The discussion was moderated by Ian Bremmer (President and Founder, Eurasia Group) and other panellists included Brad Smith (Vice-Chair and President, Microsoft); Kristalina Georgieva (Managing Director, IMF); and Khalid Al-Falih (Minister of Investment, Saudi Arabia).
“At the application layer, India will probably be the biggest supplier of services to the world,” he said, adding that return on investment (ROI) in AI comes from enterprise-level deployment and productivity gains, not from creating very large models alone.
Vaishnaw observed that nearly 95 per cent of AI use cases can be addressed with models in the 20–50 billion parameter range, many of which India already has and is deploying across sectors.
Vaishnaw, however, cautioned against equating geopolitical power with ownership of very large AI models.
He noted that such models can be switched off and may even create economic stress for their developers.
“The economics of what I call the fifth industrial revolution will come from ROI — deploying the lowest cost solution to get the highest possible return,” he said.
The minister highlighted that effective AI deployment increasingly relies on CPUs, smaller models and emerging custom silicon, reducing dependency on any single country and challenging the notion of AI dominance through scale alone.
Identifying availability of GPUs as a major constraint, he said India has adopted a public-private partnership model, empanelling around 38,000 GPUs as a common national compute facility. This facility is government-enabled and subsidised, providing affordable access to students, researchers, startups and innovators at nearly one-third the global cost.
(IANS)
