
India's AI Mission has allocated only Rs 689 crore, or 6.6% of its Rs 10,372 crore budget, to the Application Development Initiative, the layer where artificial intelligence would directly serve citizens. In…
India's AI Mission has allocated only Rs 689 crore, or 6.6% of its Rs 10,372 crore budget, to the Application Development Initiative, the layer where artificial intelligence would directly serve citizens. In contrast, compute and model development together receive about 63% of the funds, or Rs 6,534 crore, according to a SubscriberWrites piece on ThePrint.

The writer argues that India cannot outspend global players like OpenAI, which has committed $500 billion to compute alone, and should instead leverage its existing digital infrastructure, Aadhaar, UPI, and DigiLocker, to build an 'agentic layer' that can act on a citizen's behalf, such as applying for welfare schemes. The Digital Personal Data Protection Rules' consent-manager provisions, effective 13 November 2026, could provide the necessary trust layer for such agents.
The piece, which was not edited by ThePrint, suggests reprioritising funding for the application layer as shared public infrastructure, similar to how UPI was built, rather than as scattered pilots.
The writer's argument taps into a long-standing tension in Indian tech policy: building infrastructure without equivalent investment in last-mile usability. The Public Accounts Committee has documented that biometric failures in Aadhaar exclude vulnerable groups from benefits they are legally entitled to, highlighting that the system's reach does not guarantee access. An agentic layer could address this, but its success hinges on the consent-manager framework under the 2023 Digital Personal Data Protection Act, which is not yet operational. The key test will be whether the government revises its budget allocation when the mission is reviewed, or if the current spending ratios hold through the next fiscal year.
Source: theprint.in
This brief was synthesised by AI from the source linked above.