Abhishek Singh

AI Summary

→ WHAT IT COVERS Abhishek Singh, head of India's AI Mission, outlines India's $1.2 billion, five-year national AI program spanning compute infrastructure, data platforms, talent retention, and sovereign model development, positioning India as a global AI player competing with the US and China across seven strategic pillars. → KEY INSIGHTS - **Compute Subsidization Model:** India incentivizes private sector GPU investment rather than buying compute directly, then subsidizes end-user costs by 40%. This brings GPU access down to roughly $0.80 per hour versus the international rate of $2.50–$3.00 per hour, enabling IIT researchers and startups to access a pool of 38,000 GPUs through a centralized portal. - **AI Kosh Data Platform:** India's national dataset repository, AI Kosh, aggregates government and private sector data across agriculture, health, and education into one AI-ready platform. Data owners retain full access controls, choosing open or restricted sharing. Built-in anonymization tools ensure compliance with India's Digital Personal Data Protection Act before datasets are published. - **Centers of Excellence Hub-and-Spoke Structure:** India funds domain-specific AI research centers anchored at single institutions—IIT Ropar leads agriculture AI, IIT Kanpur leads smart mobility, IIT Madras leads education AI—with affiliate institutions collaborating on shared datasets and applications, replacing fragmented siloed research with coordinated national efforts across sectors. - **Brain Drain Mitigation via Ecosystem Building:** Financial grants alone do not reverse researcher emigration. Returning scientists cite the absence of venture capital networks, peer mentorship, and collaborative infrastructure as the primary barriers. India is building these support structures while also treating current US visa restrictions on H-1B holders as a strategic opportunity to attract talent back. - **Sovereign LLM Funding Strategy:** India funds at least four active foundation model development efforts, with eight more finalized for announcement. Government covers full compute costs for these projects. One effort is anchored at an IIT rather than a startup, positioning it explicitly as a state-backed sovereign model built for India's linguistic and cultural diversity. → NOTABLE MOMENT Singh reveals that as recently as last year, India had only around 500 GPUs available nationally for AI work—a figure that underscores how severe the compute gap was before the mission launched and explains why subsidized cloud access became the program's most urgent early priority. 💼 SPONSORS None detected 🏷️ India AI Policy, National Compute Infrastructure, AI Talent Retention, Sovereign AI Models, Global AI Governance