Overview
- InMobi co-founder Mohit Saxena cautioned that India will remain a large-scale user rather than a creator of AI without original IP, deep research, and system-level engineering.
- Saxena said companies are underutilising existing accelerators at roughly 7–8% and are retraining engineers on GPU programming and NVIDIA drivers to close a critical skills gap.
- NVIDIA’s Jigar Halani outlined India’s constraints of language diversity, cost sensitivity, and national-scale reliability, pointing to Bhashini and AI4Bharat as foundations for inclusive language capabilities.
- Halani argued that differentiation comes from heavy data curation and local specialisation, noting real agentic uses from Domino’s route optimisation on a single GPU to courtroom transcription and legal-assist pilots in India.
- NDTV Profit highlighted security and governance as immediate priorities, citing BCG and ManageEngine findings on widespread shadow-AI use and unapproved data entry, underscoring the need for AI-enabled cybersecurity.