The India AI Mission, backed by Rs 10,000 crore and 38,000 GPUs, picked 12 firms to build indigenous AI for social welfare. Two years in, most of these products are ready, but selling them to the very government that commissioned them is proving impossible. Companies are now pivoting their AI solutions to banks, telcos, and large enterprises instead.
How We Got Here
In March 2024, the government launched the India AI Mission, aiming for "AI of the government, for the people, by the startups." This initiative sought to apply generative AI to specific public sectors like agriculture and education, not just general chatbots.
The Numbers
- The mission intended to use AI to improve education, assist doctors, and preserve Indic languages.
- Officials asked Tech Mahindra's Project Indus team questions like "who would take over if the model suddenly glitched?"
- Companies found no fixed standards or KPIs to meet for government adoption, unlike typical contracts.
- Tech Mahindra has repurposed Project Indus as IndusQ for Qualcomm's AI hub, focusing on edge AI for smart cities.
- The government lacks a procurement framework and cross-ministry coordination to acquire these AI products.
What Happens Next
🇮🇳 Why This Matters for India
For product managers building B2G solutions in Mumbai or Hyderabad, this exposes the hard reality of government adoption beyond the grant stage.
The Take
This is a governance problem, not a tech challenge. The government built a supply-side for AI without bothering to create the demand-side, effectively subsidizing enterprise AI solutions.
Source:
The Ken ↗