India AI Impact Summit 2026, New Delhi, 16-20 February: the declaration, all 91 endorsers, seven chakras, MANAV vision, pledges and every past AI summit.
The India AI Impact Summit 2026 was held at Bharat Mandapam, New Delhi, from 16 to 20 February 2026, with the exhibition kept open for an extra day on 21 February. It was the fourth summit in the global AI series that began at Bletchley Park in 2023, and the first to be hosted by a country of the Global South. The IndiaAI Mission under the Ministry of Electronics and Information Technology (MeitY) organised it.
Prime Minister Narendra Modi opened the AI Impact Expo on 16 February and inaugurated the summit on 19 February, where he set out the M.A.N.A.V. vision for human-centric AI. French President Emmanuel Macron and UN Secretary-General António Guterres also addressed the opening ceremony.
The summit's outcome document is the AI Impact Summit Declaration, widely called the New Delhi Declaration on AI Impact. It was released on 21 February with 88 endorsers and stood at 91 by 23 February — including the United States, China, Russia, the United Kingdom and the European Union on one text. Everything in it is voluntary and non-binding.
One summit, three sets of dates
The summit week ran 16–20 February (expo to 21 February). The inauguration, Leaders' Plenary and CEO Roundtable were on 19 February. The declaration itself is headed "New Delhi, 18–19 February 2026" and says participants "gathered in New Delhi on 19 February 2026". A question may use any of the three.
PM Modi inaugurates the India AI Impact Expo 2026 at Bharat Mandapam. The 24-hour AI responsibility pledge drive begins (16–17 February).
IT Minister Ashwini Vaishnaw announces that India will add 20,000 GPUs to the existing 38,000. Awards ceremony for the three Global Impact Challenges — AI for ALL, AI by HER and YUVAi.
Research Symposium on AI and its Impact. In the evening the Prime Minister welcomes visiting leaders at Bharat Mandapam.
Summit inaugurated by PM Modi; opening ceremony also addressed by President Macron and UN Secretary-General Guterres. PM unveils the M.A.N.A.V. vision. Leaders' Plenary and, in the evening, the CEO Roundtable. Swiss President Guy Parmelin announces that Geneva will host the 2027 summit. Reliance announces its US$110 billion plan.
India signs the Pax Silica Declaration and the India–US AI Opportunity Partnership joint statement. The Global Digital Capacity Building Alliance is launched. Vaishnaw: over 5 lakh visitors, over US$250 billion pledged, 70 countries already on the declaration.
Expo extended by a day. The AI Impact Summit Declaration is released with 86 countries and 2 international organisations (88).
Bangladesh, Costa Rica and Guatemala join, taking the endorsers to 91.
The summit was built on three Sutras — guiding principles — and seven Chakras, the thematic working groups that turned those principles into deliverables. The declaration follows the same seven headings, one paragraph and one named initiative for each.
People
Sutra 1
AI should serve all of humanity, preserving dignity and ensuring inclusion.
Planet
Sutra 2
AI innovation should align with environmental stewardship and sustainability.
Progress
Sutra 3
The benefits of AI should be shared equitably to advance global development.
| Chakra (working group) | What the declaration says | Deliverable |
|---|---|---|
| Democratizing AI Resources | Affordable access to connectivity, compute, data and models; invokes Vasudhaiva Kutumbakam | Charter for the Democratic Diffusion of AI — to be run through the MAITRI platform |
| AI for Economic Growth and Social Good | Wide adoption of AI, open-source approaches where appropriate, sharing of use cases | Global AI Impact Commons — a repository of AI use cases that can be replicated across regions |
| Safe and Trusted AI | Security and trust across the AI lifecycle; industry-led voluntary measures | Trusted AI Commons and its Guidance Note on AI Governance |
| Science | Remove barriers to AI research infrastructure; international scientific collaboration | International Network of AI for Science (AI4S) Institutions |
| Inclusion for Social Empowerment | AI to widen access to knowledge, services and opportunities | Alliance for Advancing Inclusion Through AI |
| Human Capital | AI skilling, workforce development, training of public officials, AI literacy | Voluntary Guiding Principles for Skilling and Reskilling in the Age of AI; Equitable AI Transitions Playbook |
| Resilience, Innovation and Efficiency | Energy-efficient, affordable AI systems, given AI's demand for energy and resources | Voluntary Guiding Principles on Resilient, Innovative and Efficient AI; Playbook on Advancing Resilient AI Infrastructure; Resilient AI Challenge |
The official title is the AI Impact Summit Declaration; the government and the press also call it the New Delhi Declaration on AI Impact. It is a four-page text: a preamble, one section for each of the seven Chakras, and a conclusion. It creates no binding obligation — the words "voluntary" and "non-binding" recur throughout, and the initiatives are "noted" or "recognised", not adopted as commitments.
88, 89 or 91? The count depends on the date
What changed from Paris 2025
At the Paris AI Action Summit in February 2025 the United States and the United Kingdom did not sign the summit statement. In New Delhi both endorsed the declaration, alongside China and Russia. The price of that breadth was a text with no binding commitments and very little on AI safety — the theme the series began with at Bletchley Park.
The list below is reproduced from the declaration as published on the summit's official website, in the order printed there, "as on 23rd February 2026": 89 countries and two international organisations.
| No. | Endorser | Type | Note |
|---|---|---|---|
| 1 | Albania | Country | — |
| 2 | Armenia | Country | — |
| 3 | Australia | Country | Also endorsed 2 of the 6 working-group deliverables |
| 4 | Austria | Country | — |
| 5 | Bangladesh | Country | Joined after 21 February |
| 6 | Belgium | Country | — |
| 7 | Bhutan | Country | — |
| 8 | Bolivia | Country | — |
| 9 | Botswana | Country | — |
| 10 | Brazil | Country | Also endorsed 6 of the 6 working-group deliverables |
| 11 | Bulgaria | Country | — |
| 12 | Cambodia | Country | Also endorsed 6 of the 6 working-group deliverables |
| 13 | Canada | Country | Also endorsed 6 of the 6 working-group deliverables |
| 14 | Chile | Country | — |
| 15 | China | Country | — |
| 16 | Croatia | Country | — |
| 17 | Costa Rica | Country | Joined after 21 February |
| 18 | Cuba | Country | — |
| 19 | Cyprus | Country | — |
| 20 | Czech Republic | Country | — |
| 21 | Denmark | Country | — |
| 22 | Egypt | Country | Also endorsed 6 of the 6 working-group deliverables |
| 23 | Estonia | Country | — |
| 24 | Ethiopia | Country | Also endorsed 6 of the 6 working-group deliverables |
| 25 | Fiji | Country | — |
| 26 | Finland | Country | Also endorsed 2 of the 6 working-group deliverables |
| 27 | France | Country | Also endorsed 6 of the 6 working-group deliverables |
| 28 | Gambia | Country | — |
| 29 | Germany | Country | — |
| 30 | Greece | Country | — |
| 31 | Guatemala | Country | Joined after 21 February |
| 32 | Guyana | Country | — |
| 33 | Hungary | Country | — |
| 34 | Iceland | Country | Also endorsed 6 of the 6 working-group deliverables |
| 35 | India | Country | Host |
| 36 | Indonesia | Country | — |
| 37 | Iran | Country | — |
| 38 | Ireland | Country | — |
| 39 | Israel | Country | — |
| 40 | Italy | Country | Also endorsed 6 of the 6 working-group deliverables |
| 41 | Japan | Country | Also endorsed 6 of the 6 working-group deliverables |
| 42 | Kazakhstan | Country | — |
| 43 | Kenya | Country | — |
| 44 | Kyrgyzstan | Country | — |
| 45 | Latvia | Country | — |
| 46 | Liechtenstein | Country | — |
| 47 | Lithuania | Country | — |
| 48 | Luxembourg | Country | — |
| 49 | Maldives | Country | — |
| 50 | Malta | Country | — |
| 51 | Mauritius | Country | Also endorsed 6 of the 6 working-group deliverables |
| 52 | Mexico | Country | Also endorsed 6 of the 6 working-group deliverables |
| 53 | Morocco | Country | — |
| 54 | Mozambique | Country | Also endorsed 6 of the 6 working-group deliverables |
| 55 | Myanmar | Country | — |
| 56 | Nepal | Country | — |
| 57 | Netherlands | Country | Also endorsed 1 of the 6 working-group deliverables |
| 58 | New Zealand | Country | — |
| 59 | Norway | Country | Also endorsed 6 of the 6 working-group deliverables |
| 60 | Oman | Country | — |
| 61 | Paraguay | Country | — |
| 62 | Peru | Country | — |
| 63 | Philippines | Country | Also endorsed 6 of the 6 working-group deliverables |
| 64 | Poland | Country | — |
| 65 | Portugal | Country | — |
| 66 | Romania | Country | — |
| 67 | Russia | Country | — |
| 68 | Rwanda | Country | Also endorsed 1 of the 6 working-group deliverables |
| 69 | Saudi Arabia | Country | — |
| 70 | Senegal | Country | — |
| 71 | Serbia | Country | Also endorsed 6 of the 6 working-group deliverables |
| 72 | Seychelles | Country | — |
| 73 | Singapore | Country | Also endorsed 2 of the 6 working-group deliverables |
| 74 | Slovakia | Country | — |
| 75 | Slovenia | Country | — |
| 76 | South Korea | Country | — |
| 77 | Spain | Country | — |
| 78 | Sri Lanka | Country | — |
| 79 | Suriname | Country | — |
| 80 | Sweden | Country | — |
| 81 | Switzerland | Country | Also endorsed 5 of the 6 working-group deliverables |
| 82 | Tajikistan | Country | — |
| 83 | Tanzania | Country | Also endorsed 6 of the 6 working-group deliverables |
| 84 | Trinidad and Tobago | Country | — |
| 85 | UAE | Country | — |
| 86 | Ukraine | Country | — |
| 87 | UK | Country | Also endorsed 4 of the 6 working-group deliverables |
| 88 | USA | Country | — |
| 89 | Uzbekistan | Country | — |
| 90 | European Union (EU) | International organisation | — |
| 91 | International Fund for Agricultural Development (IFAD) | International organisation | — |
Reading the list
Sixteen outcome documents are published on the summit's Key Outcomes page: eleven official outcomes and five reports of Expert Engagement Groups (expert panels convened by MeitY before the summit). The five that matter most for an exam are described first.
Charter for the Democratic Diffusion of AI
Democratizing AI Resources · 24 endorsers
A voluntary framework of seven principles to widen access to compute, datasets, foundation models and talent. It will be implemented through MAITRI — Multistakeholder AI for Trusted & Resilient Infrastructure — a federated platform built as a Digital Public Good. The base MAITRI platform will be developed and maintained by the Government of India. Progress is to be reviewed annually.
Global AI Impact Commons
AI for Economic Growth and Social Good
A voluntary, sector-agnostic repository of AI use cases and success stories from across the world, so that a solution proven in one country can be adopted and scaled in another. The inclusion alliance and the science network both plug into it.
Trusted AI Commons
Safe and Trusted AI · 22 endorsers
An open repository of technical AI safety resources — benchmarks, evaluation tools, datasets, standards and best practices. India will build and maintain the base platform through a Secretariat; interested countries guide it through an informal Steering Group. Comes with a Guidance Note on AI Governance.
Network of AI for Science (AI4S) Institutions
Science · 19 endorsers
A virtual network connecting research institutions, universities and laboratories to pool AI research capabilities. It is not a funding body, regulator or standard-setter. The platform and Secretariat are hosted by India; members pay nothing to join.
New Delhi Frontier AI Impact Commitments
Industry · 13 companies
Two voluntary commitments by AI developers: (1) publish, by the next AI Summit, statistics from anonymised usage data showing how AI is actually being used in the economy; (2) evaluate their models on more languages and cultural contexts. Progress is to be reported through mechanisms such as the Hiroshima AI Process Reporting Framework.
| Document | Category | From | What it is |
|---|---|---|---|
| AI Impact Summit Declaration | Official outcome | Summit | The leaders' declaration — 91 endorsers |
| Charter for the Democratic Diffusion of AI | Official outcome | Democratizing AI Resources | Seven principles on access to AI resources; MAITRI platform |
| Framework for the Trusted AI Commons | Official outcome | Safe and Trusted AI | Repository of technical AI safety tools and benchmarks |
| Guidance Note on AI Governance | Official outcome | Safe and Trusted AI | Companion note on governance that enables adoption as well as managing risk |
| New Delhi Frontier AI Impact Commitments | Official outcome | AI companies | Usage-data transparency and multilingual evaluation |
| Network of AI for Science Institutions | Official outcome | Science | Virtual network of AI-for-science institutions |
| Alliance for Advancing Inclusion Through AI | Official outcome | Inclusion for Social Empowerment | Forum and use-case repository for inclusive AI; nine supporting principles |
| Guiding Principles for Skilling and Reskilling in the Age of AI | Official outcome | Human Capital | Voluntary principles on AI skilling and workforce readiness |
| Equitable AI Transitions Playbook | Official outcome | Human Capital | Playbook on AI workforce development |
| Guiding Principles on Resilient, Innovative and Efficient AI | Official outcome | Resilience, Innovation and Efficiency | Ten voluntary principles on frugal, energy-efficient AI |
| Playbook on Advancing Resilient AI Infrastructure | Official outcome | Resilience, Innovation and Efficiency | Reference resource on AI's energy and resource footprint |
| AI and Child Safety | Expert Engagement Group | Expert panel | Final report of the expert group on child safety |
| AI and Climate | Expert Engagement Group | Expert panel | Report on AI and climate |
| Frontier AI Model Usage Commitments | Expert Engagement Group | Expert panel | Expert input behind the Frontier AI Impact Commitments; chaired by Dr Kalika Bali |
| Human Flourishing with AI | Expert Engagement Group | Expert panel | Report on human flourishing with AI |
| New Deal for Data | Expert Engagement Group | Expert panel | Report on data for AI |
| No. | Principle |
|---|---|
| 1 | Enhance digital inclusion and connectivity for AI democratization |
| 2 | Facilitate and support representative AI development (low-resource languages, under-represented contexts) |
| 3 | Support pooled AI infrastructure initiatives (shared compute hubs, cloud pooling, federated data) |
| 4 | Encourage openness and trustworthiness in AI innovation and development |
| 5 | Inclusive and equitable benefit-sharing |
| 6 | Enhance AI capacity building and talent development |
| 7 | Support innovative financing assistance for AI infrastructure |
| No. | Organisation | Based in |
|---|---|---|
| 1 | Amazon | United States |
| 2 | Anthropic | United States |
| 3 | BharatGen | India |
| 4 | Cohere | Canada |
| 5 | United States | |
| 6 | Gnani.ai | India |
| 7 | G42 | UAE |
| 8 | Meta | United States |
| 9 | Microsoft | United States |
| 10 | Mistral AI | France |
| 11 | OpenAI | United States |
| 12 | Sarvam AI | India |
| 13 | Soket AI Labs | India |
Far fewer governments put their names to the working-group deliverables than to the declaration: between 19 and 24 endorsers each, against 91. The table is compiled from the endorser list printed at the end of each document, all "as on 23rd February 2026". India, as host and author, is not listed in any of the six. Neither the United States nor China endorsed any of them.
| Endorser | Charter (24) | Trusted AI Commons (22) | AI for Science Network (19) | Inclusion Alliance (21) | Skilling Principles (24) | Resilient AI Principles (22) | Total |
|---|---|---|---|---|---|---|---|
| Countries (25) | |||||||
| Australia | Yes | — | — | — | Yes | — | 2 |
| Brazil | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Cambodia | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Canada | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Colombia | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Egypt | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Ethiopia | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Finland | — | — | — | — | Yes | Yes | 2 |
| France | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Iceland | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Italy | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Japan | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Mauritius | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Mexico | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Mozambique | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Netherlands | — | — | — | — | — | Yes | 1 |
| Nigeria | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Norway | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Philippines | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Rwanda | — | — | — | — | Yes | — | 1 |
| Serbia | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| Singapore | — | Yes | Yes | — | — | — | 2 |
| Switzerland | Yes | Yes | — | Yes | Yes | Yes | 5 |
| Tanzania | Yes | Yes | Yes | Yes | Yes | Yes | 6 |
| UK | — | Yes | — | Yes | Yes | Yes | 4 |
| International organisations and foundations (5) | |||||||
| Asian Development Bank | Yes | — | — | — | — | — | 1 |
| The Gates Foundation | Yes | — | — | — | — | — | 1 |
| The Rockefeller Foundation | Yes | — | — | — | — | — | 1 |
| UNICEF | — | Yes | — | Yes | Yes | — | 3 |
| World Bank | Yes | — | — | — | — | — | 1 |
Three things the table shows
In his inaugural address on 19 February, Prime Minister Modi summed up India's approach to AI in the acronym M.A.N.A.V. — the Hindi word for "human". He said AI must be a "multiplier, not a monopoly", that people must not become mere data points or raw material for AI, and that the AI space should be "child-safe and family-guided". He called for authenticity labels and watermarking against deepfakes, and closed with: "Design and Develop in India. Deliver to the World. Deliver to Humanity."
| Letter | Pillar | What it means |
|---|---|---|
| M | Moral and Ethical Systems | AI rooted in ethical guidance — fairness, transparency and human oversight |
| A | Accountable Governance | Transparent rules and robust oversight |
| N | National Sovereignty | Whose data, his right — sovereignty over data, algorithms and digital infrastructure |
| A | Accessible and Inclusive AI | AI as a multiplier for society, not a monopoly of a few |
| V | Valid and Legitimate Systems | AI that is lawful, verifiable and trustworthy |
António Guterres
UN Secretary-General
Called for a US$3 billion Global Fund on AI to give developing countries basic capacity in skills, data and affordable compute — "less than one per cent of the annual revenue of a single tech company". Said the future of AI "cannot be left to the whims of a few billionaires".
Emmanuel Macron
President of France
Said protecting children from the harms of social media and AI would be a priority of France's G7 presidency in 2026; France is pursuing a ban on social media for children under 15, and he urged India to join the effort.
Michael Kratsios
Director, White House Office of Science and Technology Policy
Led the US presence. Said the United States "totally reject[s] global governance of AI" and offered partners "sovereign AI capability" built on American technology. The US still endorsed the declaration.
Guy Parmelin
President of the Swiss Confederation
Announced that Geneva will host the next summit in 2027, with a focus on international law and commonly accepted rules for AI.
The Ministry of External Affairs named 20 leaders attending at the Prime Minister's invitation, besides the UN Secretary-General. At the close, the MEA spokesperson said 20 world leaders had attended, 45 delegations were at ministerial level and about 100 countries were represented. Brazil's President Lula combined the summit with a State Visit to India (18–22 February).
| Country | Leader | Office |
|---|---|---|
| Bhutan | Tshering Tobgay | Prime Minister |
| Bolivia | Edman Lara Montaño | Vice President |
| Brazil | Luiz Inácio Lula da Silva | President |
| Croatia | Andrej Plenković | Prime Minister |
| Estonia | Alar Karis | President |
| Finland | Petteri Orpo | Prime Minister |
| France | Emmanuel Macron | President |
| Greece | Kyriakos Mitsotakis | Prime Minister |
| Guyana | Bharrat Jagdeo | Vice President |
| Kazakhstan | Olzhas Bektenov | Prime Minister |
| Liechtenstein | Hereditary Prince Alois | Hereditary Prince |
| Mauritius | Navinchandra Ramgoolam | Prime Minister |
| Netherlands | Dick Schoof | Prime Minister |
| Serbia | Aleksandar Vučić | President |
| Seychelles | Sébastien Pillay | Vice President |
| Slovakia | Peter Pellegrini | President |
| Spain | Pedro Sánchez | President of the Government (Prime Minister) |
| Sri Lanka | Anura Kumara Dissanayake | President |
| Switzerland | Guy Parmelin | President of the Swiss Confederation |
| United Arab Emirates | Sheikh Khaled bin Mohamed bin Zayed Al Nahyan | Crown Prince of Abu Dhabi |
The government's headline number is over US$250 billion in AI infrastructure pledges plus about US$20 billion of venture capital for deep tech (Ashwini Vaishnaw, 20 February). Reuters put the infrastructure total at "more than US$200 billion". These are announcements by companies, not signed contracts, and most run over five to ten years.
| Company / investor | Amount | What for |
|---|---|---|
| Reliance Industries and Jio | US$110 billion over seven years | AI and data infrastructure, including gigawatt-scale data centres |
| Adani Group | US$100 billion by 2035 | AI data centres powered by renewable energy |
| Microsoft | On track for US$50 billion by the end of the decade | AI across the Global South — not India alone; India already had a US$17.5 billion Microsoft commitment |
| US$15 billion AI hub (Visakhapatnam); two US$30 million funds | New India–US subsea cable routes; AI for Government challenge; AI for Science fund | |
| Lightspeed Venture Partners | US$10 billion | Venture investment |
| General Catalyst | US$5 billion over five years | Venture investment |
| Yotta | More than US$2 billion | AI computing hub |
| Blackstone (lead investor) | US$600 million | Equity investment in Neysa, an Indian AI cloud start-up |
| Larsen & Toubro with Nvidia | Not disclosed | Venture to build what L&T calls India's largest AI factory |
| OpenAI with the Tata Group / TCS | Not disclosed | OpenAI to be the first customer of TCS's data-centre unit, under the Stargate initiative |
| AMD with TCS | Up to 200 MW of AI capacity | Expanded partnership to deploy AI computing capacity |
| Anthropic | Not disclosed | Opened an office in Bengaluru |
| Challenge | Position | Winner | Grant |
|---|---|---|---|
| YUVAi — Global Youth Challenge | First Prize | Paraspeak | ₹15 lakh |
| YUVAi — Global Youth Challenge | First Prize | AgniSena | ₹15 lakh |
| YUVAi — Global Youth Challenge | First Prize | Z-TRACS | ₹15 lakh |
| YUVAi — Global Youth Challenge | Second Prize | CytoScanZ | ₹10 lakh |
| YUVAi — Global Youth Challenge | Second Prize | WAYV | ₹10 lakh |
| YUVAi — Global Youth Challenge | Second Prize | VOX-AID | ₹10 lakh |
| YUVAi — Global Youth Challenge | Special Recognition | MalariaX | ₹5 lakh |
| YUVAi — Global Youth Challenge | Special Recognition | KedovoAI | ₹5 lakh |
| AI by HER | Winner | Farmer Lifeline | ₹25 lakh |
| AI by HER | First Runner-up | Periwinkle Technologies | ₹25 lakh |
| AI by HER | Second Runner-up (joint) | Volar Alta | ₹25 lakh |
| AI by HER | Second Runner-up (joint) | Remidio | ₹25 lakh |
| AI for ALL | Top 10 | Infiheal Healthtech | ₹25 lakh |
| AI for ALL | Top 10 | EQUITWIN by Infiuss Health | ₹25 lakh |
| AI for ALL | Top 10 | One Global Medical Technology (Helium Health) | ₹25 lakh |
| AI for ALL | Top 10 | Resilience360 | ₹25 lakh |
| AI for ALL | Top 10 | SatSure — Farm Score for Climate-Smart Lending | ₹25 lakh |
| AI for ALL | Top 10 | Wysa — conversational AI for mental health | ₹25 lakh |
| AI for ALL | Top 10 | Kidaura Innovations | ₹25 lakh |
| AI for ALL | Top 10 | MadhuNetrAI — diabetic retinopathy screening | ₹25 lakh |
| AI for ALL | Top 10 | CarbGeM Inc. | ₹25 lakh |
| AI for ALL | Top 10 | Biome Makers — BeCrop soil intelligence platform | ₹25 lakh |
The IndiaAI Mission behind the summit
The IndiaAI Mission was approved by the Union Cabinet on 7 March 2024 with an outlay of ₹10,371.92 crore. Its seven pillars are Compute, Foundation Models, Application Development, FutureSkills, AIKosh (datasets), Safe and Trusted AI, and Startup Financing. The India AI Governance Guidelines were released on 5 November 2025.
The names tell the story of the series: Safety (2023), Seoul (2024), Action (2025), Impact (2026). Each summit moved a step away from the risks of frontier AI and towards adoption, investment and development.
| No. | Summit | Dates | Host / venue | Outcome | Endorsers |
|---|---|---|---|---|---|
| 1 | AI Safety Summit | 1–2 November 2023 | United Kingdom — Bletchley Park | Bletchley Declaration | 28 countries and the EU |
| 2 | AI Seoul Summit | 21–22 May 2024 | South Korea, co-hosted with the UK — Seoul | Seoul Declaration; Frontier AI Safety Commitments | Declaration: 10 countries and the EU; commitments: 16 companies |
| 3 | AI Action Summit | 10–11 February 2025 | France, co-chaired with India — Paris | Statement on Inclusive and Sustainable AI for People and the Planet | 58 countries at the summit; the US and the UK did not sign |
| 4 | AI Impact Summit | 16–20 February 2026 | India — Bharat Mandapam, New Delhi | AI Impact Summit Declaration (New Delhi Declaration) | 91 — 89 countries, the EU and IFAD |
| 5 | AI Summit, Geneva | 2027 | Switzerland, co-chaired with the UAE — Geneva | To be held | — |
| 6 | AI Summit | 2028 | United Arab Emirates | To be held | — |
Favourite traps
One-liners for SSC, Banking and Railway
For UPSC Prelims and Mains
It was held at Bharat Mandapam, New Delhi, from 16 to 20 February 2026, with the expo extended to 21 February. Prime Minister Narendra Modi inaugurated the summit on 19 February, the day of the Leaders' Plenary and the CEO Roundtable.
The AI Impact Summit Declaration had 88 endorsers when it was released on 21 February 2026 — 86 countries and 2 international organisations. Bangladesh, Costa Rica and Guatemala then joined, taking the total to 91 as on 23 February: 89 countries, the European Union and IFAD.
The three Sutras are People, Planet and Progress. The seven Chakras are the working groups: Human Capital; Inclusion for Social Empowerment; Safe and Trusted AI; Science; Resilience, Innovation and Efficiency; Democratizing AI Resources; and AI for Economic Growth and Social Good.
It is the human-centric AI framework Prime Minister Modi set out on 19 February 2026: Moral and Ethical Systems, Accountable Governance, National Sovereignty, Accessible and Inclusive AI, and Valid and Legitimate Systems. "Manav" means human.
Yes. Both are on the list of 91 endorsers, as are Russia, the United Kingdom and the European Union. At the Paris AI Action Summit in 2025 the United States and the United Kingdom had not signed the summit statement.
MAITRI stands for Multistakeholder AI for Trusted & Resilient Infrastructure. It is the platform through which the Charter for the Democratic Diffusion of AI will be implemented — a federated Digital Public Good for discovering affordable compute, sharing datasets and collaborating on small and local-language models. The Government of India will develop and maintain the base platform.
No. The declaration and every initiative it names — the Charter, the Global AI Impact Commons, the Trusted AI Commons, the AI for Science network and the guiding principles — are described as voluntary and non-binding.
The first was the AI Safety Summit at Bletchley Park, United Kingdom, on 1–2 November 2023, followed by Seoul (May 2024), Paris (February 2025) and New Delhi (February 2026). The next will be in Geneva, Switzerland, in 2027, co-chaired by the UAE, which hosts in 2028.