Current Affairs Topics Archive
International Relations Economics Polity & Governance Environment & Ecology Science & Technology Internal Security Geography Social Issues Art & Culture Modern History

India AI Impact Summit | Delhi declaration: Focus on democratising AI, US & China among 88 signatories

February 21, 2026 6 min read Science & Technology International Relations GS3 (Science & Technology) GS2 (International Relations)
On This Page ▼

What Happened

  • The India AI Impact Summit 2026 was held in New Delhi on February 18–19, concluding with the adoption of the New Delhi Declaration on AI Impact, endorsed by 88–91 countries and international organisations.
  • The Declaration is a voluntary, non-binding agreement centred on inclusive, human-centric AI development, with a particular focus on democratising AI access and ensuring Global South participation.
  • Notably, the United States and China did not sign the Declaration — making it a statement of "middle power" and Global South consensus rather than the great power consensus achieved at Bletchley (UK, 2023).
  • The Summit brought together over 20 Heads of Government, representatives from 118 countries, 100+ global AI CEOs, and over 5 lakh participants — one of the largest AI gatherings globally.
  • India positioned the Summit around seven "chakras" (pillars): human capital development, broadening AI access, AI trustworthiness, energy efficiency, AI for science, democratising AI resources, and AI for economic growth and social good.

Static Topic Bridges

Global AI Governance: Summit Architecture and Competing Visions

Since 2023, a series of international AI governance summits has produced a patchwork of declarations reflecting competing visions for how AI should be regulated. The Bletchley Declaration (November 2023, UK): focused on frontier AI safety risks — the dangers of very powerful AI systems — and was signed by 28 countries including the US, China, UK, EU, and India. It emphasised transparency obligations for frontier AI developers and safety testing frameworks. The Seoul Declaration (May 2024, South Korea): built on Bletchley, reaffirming frontier AI safety and adding commitments on voluntary safety testing by AI companies. The Paris AI Action Summit (February 2025): 64 countries endorsed a Declaration focused on inclusive and sustainable AI; the US and UK declined to sign. The New Delhi Declaration (February 2026): 88+ countries endorsed, focused on democratising AI and Global South inclusion; US and China did not sign.

  • Bletchley Declaration (2023): 28 signatories; focus on frontier AI safety risks; US and China both signed.
  • Seoul Declaration (2024): South Korea; reaffirmed Bletchley commitments; added voluntary safety testing.
  • Paris AI Action Summit Declaration (2025): 64 countries; inclusive AI focus; US and UK declined.
  • New Delhi Declaration (2026): 88+ countries; democratising AI; US and China not signatories.
  • A key critique of the New Delhi Declaration: it omitted any reference to frontier AI risks — unlike Bletchley and Seoul — focusing instead on AI access and economic development.
  • The Declaration is voluntary and non-binding — implementation depends on each country's political commitment.

Connection to this news: India's AI summit deliberately shifted the global AI governance conversation from the "safety-first" framing (dominant in Bletchley/Seoul, driven by Western AI safety concerns) to an "access and development-first" framing — reflecting the priorities of the Global South, which sees AI primarily as a development opportunity rather than a safety threat.

India's AI Policy Architecture and Digital Public Infrastructure

India has been building AI capability and governance infrastructure since 2018. The National Strategy for Artificial Intelligence (NITI Aayog, 2018): positioned India as an "AI Garage" for the world — using India's scale and diversity to develop AI solutions applicable globally. The IndiaAI Mission (2024): Rs 10,372 crore scheme to build sovereign AI infrastructure — 10,000+ GPU compute capacity, Indian Foundation Model development, AI datasets, and application development. India's unique strength in AI governance is its Digital Public Infrastructure (DPI) stack — Aadhaar (identity), UPI (payments), DigiLocker (documents), and ONDC (commerce) — which provides a trusted data and identity layer that can be leveraged for responsible AI deployment.

  • IndiaAI Mission: Rs 10,372 crore; announced March 2024; components: AI Compute Facility (10,000+ GPUs), IndiaAI Innovation Centre (sovereign LLM), IndiaAI Datasets Platform, IndiaAI Application Development Initiative, IndiaAI FutureSkills, IndiaAI Startup Financing.
  • India's DPI stack: Aadhaar (1.4+ billion enrolled), UPI (12+ billion monthly transactions), DigiLocker (250+ million users).
  • IndiaAI Mission's IndiaAI Language Model: targeting development of large language models (LLMs) in India's 22 scheduled languages — directly addressing Global South language inclusion.
  • AI standardisation: India has been active in ISO/IEC AI standards bodies; pushed for AI governance through G20 (India held G20 Presidency in 2023).
  • G20 New Delhi Declaration (2023): included AI principles on safe, trustworthy AI; India authored the AI governance section.
  • GPAI (Global Partnership on AI): India is a founding member of this multilateral initiative for responsible AI development.

Connection to this news: India's positioning at the AI Impact Summit is coherent with its existing AI policy architecture — the New Delhi Declaration's emphasis on democratising AI access and Global South LLMs directly mirrors the IndiaAI Mission's domestic objectives, allowing India to be a rule-setter internationally based on what it is already building domestically.

Geopolitics of AI: US-China Competition and India's Strategic Positioning

Artificial intelligence has become the central arena of US-China great power competition, with profound implications for military capability, economic productivity, and technological leadership. The US has imposed sweeping export controls on advanced AI chips (particularly Nvidia's H100 and A100 GPUs) to China, attempting to deny Chinese AI developers access to the best training hardware. China has responded by accelerating domestic chip development (Huawei's Ascend series) and AI model development (DeepSeek's R1 model, released January 2025, demonstrated competitive performance with US frontier models at a fraction of the cost). The absence of both the US and China from the New Delhi Declaration reflects this bifurcation: neither great power wants to be constrained by a multilateral framework that the other might use strategically.

  • Nvidia export controls: US Bureau of Industry and Security (BIS) has progressively tightened controls on exporting A100, H100, and successor GPU chips to China since 2022.
  • DeepSeek R1 (January 2025): Chinese AI model that matched GPT-4 level performance at ~$6 million training cost vs. estimated $100 million+ for comparable US models — demonstrated that compute controls may not stop Chinese AI progress.
  • AI compute: measured in FLOPs (floating point operations); frontier model training requires 10^24–10^26 FLOPs.
  • India's AI compute gap: India's current GPU capacity is a fraction of US and Chinese capability; IndiaAI Mission's 10,000+ GPU target is a starting point.
  • India-US iCET (Initiative on Critical and Emerging Technologies): covers AI, quantum computing, semiconductors, space, and advanced telecom — India is a strategic partner for US in AI ecosystem.
  • The New Delhi Declaration's focus on inclusive AI governance aligns with UN AI Advisory Body's recommendations (2024) for a global AI framework that serves developing nations.

Connection to this news: India's hosting of the AI Impact Summit — and the 88-country endorsement of the New Delhi Declaration — positions India as the leader of a "third path" in global AI governance: neither the US-led "safety-first" restrictive approach nor the Chinese state-directed approach, but an inclusive, development-oriented framework that speaks for the Global South.

Key Facts & Data

  • India AI Impact Summit 2026: New Delhi, February 18–19.
  • New Delhi Declaration: endorsed by 88–91 countries and international organisations.
  • US and China: did not sign the New Delhi Declaration.
  • Summit participants: 20+ Heads of Government, representatives from 118 countries, 100+ AI CEOs, 5 lakh+ total participants.
  • Seven Chakras (pillars): human capital, AI access for social empowerment, trustworthiness, energy efficiency, AI for science, democratising AI resources, AI for growth and social good.
  • Bletchley Declaration (2023): 28 countries, frontier AI safety focus; US and China both signed.
  • Paris AI Summit Declaration (2025): 64 countries; US and UK declined.
  • New Delhi Declaration: 88+ countries; largest signatory count of any AI summit declaration to date.
  • IndiaAI Mission: Rs 10,372 crore (2024); includes 10,000+ GPU compute, sovereign LLM, 22-language Indian LLM.
  • G20 AI governance (2023): India authored AI principles section during India G20 Presidency.
  • GPAI (Global Partnership on AI): India is a founding member.
  • India's DPI stack: Aadhaar (1.4 billion), UPI (12+ billion monthly transactions), DigiLocker (250+ million users).
  • DeepSeek R1 (January 2025): Chinese model matching GPT-4 at ~$6 million training cost vs $100 million+ for US equivalents.
  • India-US iCET: covers AI, quantum, semiconductors, space, advanced telecom.