New Delhi Declaration on AI Impact

Summit: India‑AI Impact Summit 2026 (24‑26 Feb 2026)\
Signatories: 89 countries & international organisations (including USA, China)\
Declaration Principle: Sarvajan Hitaya, Sarvajan Sukhaya (welfare of all)\
Seven ‘Chakras’: Framework covering democratization, trusted AI, research collaboration, social empowerment, human capital, resilience & innovation\
Investment Pledges: > USD 250 billion for data‑centres & semiconductor fabs; USD 20 billion for deep‑tech VC\
Compute Expansion: Existing GPUs 38,000 → +20,000 GPUs; target >100,000 GPUs by end‑2026\
Indigenous Models: Sarvam‑30B (30 B parameters) & Sarvam‑105B (105 B parameters); BharatGen Param2 (17 B, 22 Indian languages)\
Strategic Alliances: Joined US‑led Pax Silica (semiconductor security) and signed India‑US AI Opportunity Partnership\
Policy Initiatives: IndiaAI Mission 2.0, MANAV Vision (Moral, Accountable, National sovereignty, Accessible, Valid), UPI‑style AI Playbook for MSMEs\

The declaration builds on India’s Digital Public Infrastructure (DPI) – UIDAI, UPI, DigiLocker – positioning it as a global model for inclusive digital services.\
It reflects a shift from AI safety‑centric regimes (e.g., EU AI Act) to a democratic diffusion approach, aiming to bridge the digital divide, especially for the Global South.\
The summit coincided with rising geopolitical competition over semiconductor supply chains, prompting the Pax Silica alliance.

Sovereign AI: Indigenous foundation models reduce dependence on foreign platforms, safeguard data sovereignty and enhance strategic autonomy.\
Economic Growth: Massive GPU procurement and semiconductor investments are expected to create a $250 bn AI ecosystem, boosting MSME productivity via the AI Playbook.\
Human Capital: Skilling & reskilling initiatives under the MANAV Vision aim to up‑skill the 5.8 million IT workforce, mitigating displacement risks.\
Sustainable AI: Voluntary Guiding Principles on Resilient, Innovative, Efficient AI push for green data‑centres and energy‑efficient algorithms.

Article 19(1)(a) & (g): Freedom of speech and the right to practice any profession – supports open AI innovation while requiring reasonable restrictions for security.\
Data Protection Bill (draft): Aligns with the declaration’s emphasis on trusted AI and data sovereignty.\
National Security Act: Provides a legal basis for protecting critical semiconductor infrastructure under the Pax Silica framework.

Non‑binding nature: Voluntary commitments lack enforcement mechanisms; no global verification body.\
Absence of red‑lines: Unlike the EU AI Act, the declaration does not prohibit high‑risk AI applications (e.g., predictive policing).\
Infrastructure gaps: Current compute capacity and semiconductor fab ecosystem are insufficient for competing with US/China.\
Environmental impact: Data‑centre water consumption (~11 lakh litres/day) raises sustainability concerns.

Operationalise the Three Sutras (People, Planet, Progress): Embed MANAV Vision in all public‑sector AI projects.\
Scale the Reskilling Playbook: Partner with industry to certify AI‑ready talent for MSMEs.\
Institutionalise Trusted AI Commons: Create a multilateral technical standards body to move beyond voluntary pledges.\
Promote Green AI: Incentivise renewable‑powered data‑centres and low‑power algorithms.\
Strengthen Semiconductor Mission (ISM 2.0): Accelerate indigenous chip design and fab capacity.

Potential UPSC Questions\

Mains: Discuss how the New Delhi Declaration shifts AI governance from safety‑centric to democratic diffusion and its implications for the Global South.\
Prelims: Which of the following were pledged at the India‑AI Impact Summit 2026? (a) USD 250 bn for AI infrastructure (b) Creation of Sarvam‑105B model (c) Joining Pax Silica alliance (d) All of the above.