AI Impact Summit 2026 l Keynote Speakers

Contents

India AI Impact Summit 2026 — Complete Opening: Master Summary

Executive Summary

The India AI Impact Summit 2026, held at Bharat Mandapam in New Delhi, was the first major international AI summit hosted in the Global South and the largest AI summit ever convened, drawing participation from 118 countries across an inaugural opening session and a marathon seven-hour main keynote featuring over 30 heads of state, Nobel laureates, AI company CEOs, venture capitalists, and senior government ministers. The event produced combined investment commitments exceeding $285 billion — led by Reliance/Jio's ~$120 billion sovereign AI pledge, the Adani Group's $100 billion green-energy AI commitment, Google's ~$15 billion Visakhapatnam hub, and Microsoft's $50 billion Global South allocation — while launching India's sovereign multilingual AI models, signing the New Delhi Frontier AI Impact Commitments with frontier and Indian AI companies, and generating a remarkably unified thesis: that AI's greatest test and greatest opportunity lie not in frontier laboratories but in delivering intelligence at population scale to the billions of the Global South. Beneath this consensus, however, the summit surfaced deep, unresolved tensions — on AGI timelines (Sam Altman predicting superintelligence by end of 2028 versus Yann LeCun dismissing AGI as "a false idea"), on open-source versus proprietary models, on sovereignty versus concentration, and on whether AI will create or destroy jobs — framing these not as abstract debates but as concrete policy and investment choices that will determine whether AI narrows or widens global inequality.

Key Takeaways

India declared itself a co-architect of the AI era, not a passive consumer of imported intelligence. From Prime Minister Modi's inaugural "AI for everyone" address through Minister Vaishnaw's five-layer AI stack framework, Mukesh Ambani's "India cannot afford to rent intelligence," and Nandan Nilekani's "India needs AI and AI needs India," the summit's central narrative was that India's structurally unique assets — Aadhaar, UPI, Bhashini, 1.4 billion digital identities, 20 billion monthly payment transactions, 500 million digital health IDs — position it to define how AI reaches the world's majority populations. India launched sovereign multilingual models during the summit, created a public compute platform of 38,000 GPUs (expanding by 20,000), and attracted investment pledges that reframe AI infrastructure as nation-building capital on par with energy grids, highways, and ports.
Combined investment commitments exceeded $285 billion, signaling that AI compute is the new strategic infrastructure. Reliance/Jio pledged approximately $120 billion over seven years for sovereign AI infrastructure including gigawatt-scale data centers; the Adani Group committed $100 billion for green-energy-powered AI compute triggering a 5-GW, $250 billion integrated ecosystem; Google announced a roughly $15 billion full-stack AI hub in Visakhapatnam with gigawatt-scale compute and a new submarine cable gateway; Microsoft committed $50 billion for the Global South by decade's end; and Tata Group unveiled India's first 100-MW AI-optimized data center in partnership with OpenAI with a roadmap to 1 GW. These figures collectively repositioned AI infrastructure investment alongside traditional national infrastructure priorities such as energy, transportation, and telecommunications.
Sovereign AI infrastructure became the dominant strategic framework across industry, government, and international partners. Ambani's Jio Intelligence platform, Jeet Adani's three-pillar sovereignty model (energy, compute, services), Mistral CEO Arthur Mensch's warnings about 3–4 companies owning global AI deployment, France's €200 billion European AI commitment, and Minister Vaishnaw's insistence on compute as a public good all converged on the argument that nations must own their AI compute, data, and deployment stacks or face structural dependency. Macron encapsulated the principle: "No country is bound to serve only as a market where foreign companies sell their models and download citizens' data."
Multilingual, inclusive AI emerged as the defining bottleneck between AI capability and AI impact. India's 22 scheduled languages, Togo's 42 dialects, Indonesia's 17,000-island archipelago, and billions of non-English speakers worldwide made local-language AI the connective thread across nearly every address. Anthropic announced Claude Sonnet 4.6 with improved performance across 10 Indic languages; Meta open-sourced omnilingual models covering 1,600+ languages; France and India announced an open hardware translation tool for Indian languages; and the New Delhi Frontier AI Impact Commitments specifically targeted strengthening multilingual and contextual evaluations for the Global South.
Healthcare, education, and agriculture are AI's "holy trinity" for development — and concrete deployments are already operational. Vinod Khosla's call for AI doctors, tutors, and agronomists; the Gates Foundation's "Horizon 1000" initiative deploying AI across 1,000 clinics in Africa; Jio's six launched applications (Shikshak, Arogya, Krishi, Bharat IQ, Frames, Jio-Hotstar AI); Google's AI weather forecasts sent to millions of Indian farmers via the Neural GCM model; Anthropic's Claude+OpenAgroNet for rural advisory; Philips' autonomous MRI scanning; Wipro's TB-detection pilot in Tamil Nadu; and Wadhwani AI's reading assessment tool at 5 paise per child for 6 million children demonstrated that AI-for-development is no longer aspirational — it is operational and measurable.
The timeline to superintelligence and AGI is the AI community's most consequential and most contested question. Sam Altman predicted early superintelligence by end of 2028; Demis Hassabis estimated AGI within five years with impact "10 times the industrial revolution at 10 times the speed"; Dario Amodei described "only a small number of years" before AI surpasses most human cognitive capabilities; and Yann LeCun flatly rejected the entire framing, arguing current LLMs are "mostly information retrieval systems" missing world models for physical reasoning, that scientists have made the same 10-year prediction "four or five times in the last 70 years," and that "we don't even have self-driving cars that can teach themselves to drive in 20 hours like any 17-year-old." This fundamental disagreement — playing out on the same stage before the same audience — underscored that the pace at which societies must prepare remains profoundly uncertain, with massive implications for policy timing.
Trust, not capability, is the binding constraint on AI adoption at scale — and the Global South faces compounding asymmetries. ServiceNow's 55x growth in agentic workflows only after security controls were introduced, Kyndryl's finding that 75% of Indian organizations stall after proof-of-concept, Costa Rica's revelation that it invests only 0.3% of GDP in R&D against an OECD average of 2.7%, and UN Secretary-General Guterres's call for a $3 billion Global Fund for AI capacity in developing countries together painted a picture in which the gap between AI demos and AI deployment is fundamentally a trust, readiness, and resource gap — not a technology gap.
Edge AI and frugal inference emerged as essential for "AI for everyone." NXP's Lars Reger demonstrated a 10-billion-parameter LLM running on a dongle at 7 watts. Yann LeCun predicted AI's future lies in edge computing for the physical world. Macron praised India's strategy of task-specific Small Language Models designed to run on smartphones. Vaishnaw stated that more than 90% of use cases are addressable by smaller, cost-efficient models. Together, these contributions argued that PM Modi's vision requires moving inference to devices — not just scaling data centers — with predictions of 50 billion smart connected devices within a decade.
The open-source versus proprietary debate became a geopolitical fault line with no resolution. Mistral's Arthur Mensch warned that 3–4 companies owning AI deployment threatens global stability. Meta touted open-sourced omnilingual models covering 1,600+ languages. Yann LeCun argued open source is the natural evolution of internet infrastructure. Sunil Mittal pressed leading developers on whether AI would remain accessible or become more controlled. Meanwhile, Sam Altman warned about open-source bioweapons risk as a genuine exception to the broad-access principle. The tension between democratized access and safety-driven control ran through the entire summit without consensus.
A nascent but fragile international AI governance architecture is taking shape. UN Secretary-General Guterres announced a 40-member Independent International Scientific Panel on AI, the Global Dialogue on AI Governance (first session in Geneva, July 2026), and called for a $3 billion Global Fund for AI. Sam Altman called for an IAEA-like body. Emmanuel Macron outlined France's G7 Presidency priorities on inclusive AI and children's protection. The US White House adviser described a three-pillar approach of infrastructure, innovation, and allied technology sharing. India and France launched the International Challenge for Sustainable AI Models and the Coalition for Sustainable AI reached 200+ supporters. But the diversity of approaches — from Guterres's multilateral frameworks to the US emphasis on executive orders against onerous regulation — underscored that no single governance framework yet commands consensus, even as the technology accelerates beneath all of them.

Speaker Highlights

Narendra Modi, Prime Minister of India

Delivered the inaugural address establishing the summit's overarching framing of "AI for everyone" — a phrase that became the recurring touchstone across both sessions
Articulated the MANAV framework for ethical AI — emphasizing human vision as ethical guidance and accountable governance
Announced major policy shifts during the Union Budget to attract global data to India for processing and high-value service delivery
Referenced India's nuclear energy reform and clean energy trajectory as foundational to AI infrastructure; his address was translated in real-time into sign language by AI — cited as a landmark demonstration of AI accessibility
Notable quotes:
- (Paraphrased) "India will attract the world's data — to reside here, be processed here, and generate high-value services for the world."

Ashwini Vaishnaw, Minister for Electronics and Information Technology, Government of India

Articulated India's comprehensive five-layer AI stack strategy: applications, models/sovereign layer, compute, infrastructure, and energy
Launched India's sovereign multilingual, multimodal AI models during the summit, stating they "have stood their ground against frontier models on the points that matter most to us"
Announced the 38,000-GPU common compute platform (expanding by 20,000) offered at affordable rates as a public good for startups, academia, researchers, and students
Presided over the formal signing of the New Delhi Frontier AI Impact Commitments — two voluntary commitments on (1) advancing understanding of real-world AI usage through anonymized insights, and (2) strengthening multilingual and contextual evaluations for the Global South
Cited Stanford AI Index recognition of India among leading countries in AI adoption, talent, and diffusion; noted over 50% of India's power generation capacity from clean/renewable sources
Notable quotes:
- "India treats compute as a public good."
- "More than 90% of use cases are addressable by smaller, cost-efficient models rather than frontier systems."

Mukesh Dhirubhai Ambani, Chairman & Managing Director, Reliance Industries / Jio

Delivered the summit's most consequential financial announcement: ₹10 lakh crore (approximately $120 billion) over seven years for sovereign AI infrastructure
Launched Jio Intelligence — a comprehensive sovereign AI platform with gigawatt-scale data centers at Jamnagar (120+ MW operational in H2 2026), up to 10 GW of green energy surplus, nationwide edge compute, and six AI applications already deployed: Jio Shikshak (education), Jio Arogya (health), Jio Krishi (agriculture), Jio Bharat IQ (knowledge), Jio Frames (creative), and Jio-Hotstar AI
Articulated five non-negotiable principles: deep-tech productivity, multilingual capability across all Indian languages, data residency and trust by design, job creation over job destruction, and a partnership ecosystem with Indian academia and startups
Notable quotes:
- "India cannot afford to rent intelligence."
- "AI is a modern-day Akshaya Patra."

Sam Altman, CEO, OpenAI

Revealed 100 million+ weekly ChatGPT users in India — over a third students — and India as OpenAI's fastest-growing Codex market
Predicted early superintelligence by end of 2028, stating "more of the world's intellectual capacity could reside in data centers than outside them"
Advocated for democratized AI access as both a moral imperative and a safety strategy, arguing that concentration of AI power is inherently more dangerous than broad distribution — while warning about open-source bioweapons risk as a genuine exception
Called for an IAEA-like international AI body for governance and safety oversight; confirmed partnership with Tata Group for India's first major AI data center
Notable quotes:
- "More of the world's intellectual capacity could reside in data centers than outside them [by end of 2028]."
- "Democratized AI access is both a moral imperative and a safety strategy."

Dario Amodei, CEO, Anthropic

Spoke in both the opening session and the main keynote; described the exponential AI trajectory as "Moore's Law for intelligence" with "only a small number of years" before AI models surpass most human cognitive capabilities for most tasks
Coined the metaphor of "a country of geniuses in a data center" — AI agents more capable than most humans at most things, coordinating at superhuman speed
Announced Anthropic's new Bengaluru office with Ireena Vittal (formerly Ireena Ghosh) as Managing Director; launched enterprise partnerships with Infosys and nonprofit partnerships with EkStep Foundation, Pratham, Central Square Foundation, Karya, and the Collective Intelligence Project
Announced Claude Sonnet 4.6 with improved performance across 10 Indic languages, Claude+OpenAgroNet integration for rural farmer advisory, and reported Claude usage in India doubled in four months
Speculated that AI could enable India to achieve 20–25% economic growth rates — a number he acknowledged as "absurd" but argued was plausible when all factors are stacked
Joined the New Delhi Frontier AI Impact Commitments and announced the Anthropic Economic Index and Anthropic Economic Futures Program for sharing AI-and-jobs data with the Indian government
Notable quotes:
- "A country of geniuses in a data center."
- "AI has been on an exponential for 10 years and there are only a small number of years before AI models surpass the cognitive capabilities of most humans for most tasks."
- "India could theoretically achieve 20–25% economic growth rates."

Sundar Pichai, CEO, Alphabet and Google

Announced a full-stack AI hub in Visakhapatnam (Vizag) as part of a roughly $15 billion infrastructure investment in India, featuring gigawatt-scale compute and a new international submarine cable gateway
Launched the America India Connect Initiative — four new submarine fiber optic cable systems between the US and India
Highlighted AlphaFold (Nobel Prize-winning protein structure prediction, open database used by 3 million+ researchers in 190+ countries), AI-powered flood forecasts sent to millions of Indian farmers via Google's Neural GCM model, AI-powered diagnosis in El Salvador, expanded research across 20+ African languages in Ghana, SynthID for content verification, and AI for electrification mapping in Uganda
Announced the Google AI Professional Certificate and cited 100 million people already trained in digital skills globally
Notable quotes:
- "We cannot allow the digital divide to become an AI divide."
- "For the first time last summer, the Indian government sent AI-powered flood forecasts to millions of farmers."

Sir Demis Hassabis, Co-founder & CEO, Google DeepMind; Nobel Laureate

Described AGI's potential impact as "10 times the industrial revolution at 10 times the speed" and estimated arrival within five years
Cited AlphaFold as proof that AI can solve grand scientific challenges — compressing decades of protein-structure research into a tool used by over 3 million researchers for malaria vaccines, antibiotic resistance, and more
Announced a Reliance Jio partnership to bring Gemini foundation models to India
Cautioned against leaving AI governance solely to technologists, advocating for broad societal participation
Notable quotes:
- "10 times the industrial revolution at 10 times the speed."
- "AlphaFold's open database is used by over 3 million researchers in more than 190 countries."

Yann LeCun, Executive Chairman, Advanced Machine Intelligence Labs; "Godfather of Deep Learning"

Delivered the summit's sharpest intellectual dissent: current LLMs are "mostly information retrieval systems," AGI is a "false idea," and AI fundamentally lacks world models for physical reasoning — an illustration of Moravec's Paradox
Pointed out that scientists have predicted AGI within 10 years "four or five times in the last 70 years" and that "we don't even have self-driving cars that can teach themselves to drive in 20 hours like any 17-year-old"
Predicted AI's future lies in edge computing for the physical world, that inference cost is the primary barrier for India's population, and that long-term AI leadership will come from India and Africa due to demographics
Argued passionately for open-source AI as the natural evolution of internet infrastructure
Notable quotes:
- "LLMs are mostly information retrieval systems."
- "AGI is a false idea."
- "We don't even have self-driving cars that can teach themselves to drive in 20 hours like any 17-year-old."
- "Scientists have made the same 10-year prediction four or five times in the last 70 years."

N. Chandrasekaran, Chairman, Tata Group

Made five major announcements: India's first large-scale AI-optimized data center; a 100-MW AI data center with OpenAI (scaling to 1 GW); a partnership with AMD for sustainable high-density AI compute capacity; an AI data insights platform for diverse Indian contexts; and an agentic AI operating system for industries via TCS and Tata Communications
Announced Tata's entry into domain-centric, AI-optimized semiconductor chips targeting the automotive sector first, enabled by India's SemCon India policy reforms
Referenced a demonstration at Bharat Mandapam where 500 rural women with no computing background learned AI and built products, marketing materials, and campaigns in four hours
Notable quotes:
- "We will build domain-centric chips optimized for specific industries — automotive first."
- "500 women with no computing background built AI-powered products in four hours."

António Guterres, Secretary-General, United Nations

Announced three major governance initiatives: a 40-member Independent International Scientific Panel on AI, the Global Dialogue on AI Governance (first session in Geneva, July 2026), and a call for a $3 billion Global Fund for AI to build basic capacity in developing countries
Announced formation of a Global Network for Exchange and Cooperation on AI Capacity Building in the developing world
Cautioned about AI's soaring energy and water demands, demanding data centers switch to clean power; called for the protection of children and investment in workers so AI augments rather than only replaces human potential
Notable quotes:
- "The future of AI cannot be decided by a handful of countries or left to the whims of a few billionaires."
- "No child should be a test subject for unregulated AI."
- "Dignity as the default setting."

Emmanuel Macron, President of France

Praised India Stack as "a civilization story" — 1.4 billion digital identities, 20 billion payment transactions per month, 500 million digital health IDs — and framed the summit within Paris-to-New Delhi continuity
Detailed France's AI delivery: Mistral AI (€12 billion valuation), €58 billion in data centers powered by nuclear energy, 1,100+ AI startups, quantum computing investments in Pasqal, Quandela, Alice & Bob, Quobly; at European level, €200 billion committed including the LUMI/Alisoc exascale supercomputer
Recognized India's complementary approach of task-specific Small Language Models versus Europe's scaled LLMs; launched the Indo-French Institute for AI in Health (Sorbonne Brain Institute–AIIMS Delhi; Inria–Sanjivini Research Institute in Bangalore)
Announced an open hardware translation tool for Indian languages, the International Challenge for Sustainable AI Models (with India and UNESCO), and France's process to ban social media for children under 15 — inviting India to join
Notable quotes:
- "India Stack is a civilization story."
- "India chose granular, smart sovereign SLMs. Europe chose sovereign and scaled LLMs. Both chose independence."
- "No country is bound to serve only as a market where foreign companies sell their models and download citizens' data."

Nandan Nilekani, Co-founder, Infosys; Architect of Aadhaar

Launched "100 Diffusion Pathways by 2030" — a global coalition with Anthropic, Google, Gates Foundation, UNDP, and Kenya creating packaged, replicable playbooks for AI deployment modeled on India's DPI expansion now running in approximately 40 countries
Demonstrated dramatic implementation compression: Maharashtra's agri-stack took 9 months; the same learning applied in Ethiopia took 3 months; with Amul, 3 weeks
Advocated for massive compute investment, model imports combined with sovereign capability, language inclusion, and agents as tools for inclusion
Notable quotes:
- "India needs AI and AI needs India."

Rishi Sunak, Former Prime Minister, United Kingdom

Positioned India as the potential "Dutch Republic" of the AI era — extracting maximum value from transformative technology regardless of where it was invented
Cited India as the second-largest contributor to AI projects on GitHub, Stanford's ranking of India overtaking the UK in global AI power, and approximately 90% Indian optimism about AI
Referenced global workforce gaps — 11 million health workers, 44 million teachers, $4 trillion SDG funding gap by 2030 — as the problems AI must solve
Notable quotes:
- "AI will deliver twice the impact of the Industrial Revolution in half the time."
- "India could be the Dutch Republic of the AI era."

Brad Smith, Vice Chair and President, Microsoft

Announced $50 billion in Microsoft investment by decade's end to bring AI to the Global South, with India as a major recipient
Launched Microsoft Elevate for Educators — a global initiative to equip teachers for AI-integrated education
Drew the electricity analogy: AI, like electricity 144 years ago, is a general-purpose technology — but electricity still hasn't reached 700 million people, warning that AI could repeat the same exclusion without intentional effort
Called for clear goals, common measurement, and annual accountability across AI summits
Notable quotes:
- "144 years after the invention of electricity, 700 million people still don't have it."

Vinod Khosla, Founder, Khosla Ventures

Focused exclusively on immediate, practical applications for India's bottom half: AI personal tutors, AI doctors providing 24/7 primary care, and AI agronomists
Cited CK12.org as a working example — approximately 400 million students worldwide, 4 million in India, CBSE/NEP-compatible, available in English, Hindi, Odiya, and Meghalayan
Argued these three applications alone could transform India's human development trajectory within 1–2 years
Notable quotes:
- "AI doctors, AI tutors, and AI agronomists — these three things alone can transform India's bottom half within one to two years."

Jeet Adani, Director, Adani Digital Labs

Articulated a three-pillar framework for AI sovereignty: energy, compute/cloud, and services
Announced the Adani Group's $100 billion commitment to sovereign, green-energy-powered AI infrastructure, triggering a 5-GW, $250 billion integrated energy-and-compute ecosystem
Notable quotes:
- "AI must become a force multiplier for Indian citizens before it becomes a margin multiplier for others."

Roshni Nadar Malhotra, Chairperson, HCL Technologies

Declared "knowledge itself is being commoditized" and called for India's transition from tech-services-led to IP-led growth
Announced HCL's AI Factory and Physical AI service lines
Articulated three national imperatives: scale-led to IP-led, adoption to building, and compute as digital public infrastructure
Notable quotes:
- "Knowledge itself is being commoditized."

Shantanu Narayen, CEO, Adobe

Championed content authenticity and provenance/watermarking as paramount, claiming PM Modi personally endorsed universal content provenance
Argued AI's sustainable competitive advantage lies in use cases, not models — and cited PDF as proof that open standards win mass adoption
Notable quotes:
- "AI's sustainable competitive advantage lies in use cases, not models."

Rishad Premji, Executive Chairman, Wipro

Cited approximately 650,000 AI professionals in India projected to double by 2027, 4,000+ deep-tech AI startups, and a government initiative to train 10 million young people
Highlighted the Azim Premji Foundation's TB detection pilot in Tamil Nadu — portable X-ray devices with instant AI analysis by community health workers
Notable quotes:
- "India's 650,000 AI professionals are projected to double by 2027."

Julie Sweet, Chair and CEO, Accenture

Pushed back on job destruction narratives with data: Accenture grew from 275,000 to 750,000+ employees during the period when the 2013 Oxford study predicted 47% of US jobs were automatable
Reported 78% of C-suite leaders identify growth, not cost-cutting, as AI's greatest value
Advocated for "humans in the lead, not humans in the loop" and called for global AI safety standards alongside industry-specific standards
Notable quotes:
- "Humans in the lead, not humans in the loop."

Rajesh Subramanyan, CEO, FedEx

Framed intelligence as infrastructure: FedEx generates 2 petabytes of data daily, handles $2 trillion in goods annually, and moves 17 million packages daily
Announced the FedEx Import Tool — developed first in India for SMEs, featuring predictive logistics and AI-powered customs — now rolling out globally
Notable quotes:
- "If you don't like change, you will hate extinction."

Nikesh Arora, CEO, Palo Alto Networks

Warned AI is accelerating faster than institutions, governance, and human intuition
Identified governance, human/social impact, and cybersecurity as three core challenges
Predicted five times more technology workers will be needed; cautioned against overestimating the speed of job replacement, citing Google's self-driving car prototype taking 16 years from demo to near-functional deployment
Notable quotes:
- "In about six months if I'm talking to my AI model, it might know more things about me than I've told my wife."

Arthur Mensch, CEO, Mistral AI

Warned against a world where 3–4 companies own AI deployment and infrastructure, framing concentration as a threat to global stability
Advocated for open source as foundational infrastructure; noted approximately 25% of Mistral's researchers are Indian
Urged governments to invest in infrastructure they own and businesses to internalize AI value chains
Notable quotes:
- "A world where 3–4 companies own AI deployment threatens global stability."

Ankur Vora, Gates Foundation

Announced "Horizon 1000" (1,000 AI-enabled clinics across Africa with OpenAI and Rwanda) and "Advantage India for AI"
Cited Wadhwani AI's reading assessment tool (5 paise per child, 6 million children served) and a farmer in Andhra Pradesh using AI for drone-based pest treatment within 48 hours
Notable quotes:
- "AI could possibly compress progress of the next 20 years into five."

Vishal Sikka, Founder & CEO, VNI

Demonstrated a 250x productivity improvement from generative AI coding tools — a service rebuilt by one person in 14 days that originally required 15 engineers over 9 months
Introduced the "jagged frontier" concept — AI effectiveness is non-uniform — and identified the gap between LLMs and enterprise users as the key value-creation opportunity
Notable quotes:
- "250 times productivity improvement — one person in 14 days versus 15 engineers over 9 months."

Roy Jakobs, CEO, Philips

Described autonomous MRI scanning, AI-drafted clinical documentation, smart healing environments, and agentic AI in healthcare — all under human oversight
Cited 76% of Indian healthcare professionals believing AI can improve patient outcomes; referenced Philips' 97-year presence in India
Notable quotes:
- "76% of Indian healthcare professionals believe AI can improve patient outcomes."

Lars Reger, CTO, NXP Semiconductors

Demonstrated a 10-billion-parameter LLM running on a dongle at 7 watts — a physical embodiment of edge AI
Predicted 50 billion smart connected devices in a decade; demonstrated car-to-car communication (1 mile in 3 milliseconds) and in-car microphones detecting bicycle bells for road safety
Announced NXP's acquisition of Indian AI accelerator company Kinara
Notable quotes:
- "50 billion smart connected devices within a decade."

Amit Zavery, President & CPO, ServiceNow

Positioned trust as the new infrastructure for enterprise AI; revealed 55x growth in agentic workflow adoption only after security controls were introduced
Argued 80–90% of enterprise AI IP comes from the surrounding software layer, not foundational models
Announced acquisition of "VZA" for non-human identity management
Notable quotes:
- "Trust is the new infrastructure for enterprise AI."

Martin Schroeter, CEO, Kyndryl

Diagnosed the core AI challenge as readiness, not innovation: 75% of Indian organizations stall after proof-of-concept; nearly half of globally invested organizations struggle to see returns
Highlighted the Unified Lending Interface (ULI), Bangalore Airport's self-healing AI operations, and a new cyber defense operations center opening in India
Notable quotes:
- "75% of Indian organizations stall after proof-of-concept."

Sunil Mittal, Founder/Chairman, Bharti Airtel

Pressed leading AI developers on the open-source question — whether AI would remain accessible or become more controlled — surfacing a critical tension running through the entire summit

Bejul Somaia, Managing Director, Lightspeed Venture Partners

Drew the parallel between India's 2008 digital skepticism (low-single-digit internet penetration) and its emergence as the world's third-largest digital economy
Notable quotes:
- "Scale is a snapshot, but slope and trajectory are the story."

Olivier Blum, CEO, Schneider Electric

Termed the era "shadow electric energy intelligence" and argued India is at an inflection point where it can innovate rather than import
Notable quotes:
- "If you can crack the code in India, we'll crack the code everywhere."

Sriram Krishnan, Senior Policy Adviser for AI, White House

Outlined the Trump administration's three AI priorities: infrastructure, innovation (including an executive order against onerous regulation), and technology sharing with allies
Referenced the AI Accelerator Partnership with UAE and the Pax Silica supply-chain security initiative

Ministerial Panel Speakers

Cina Lawson (Togo): Shared pandemic-era satellite poverty mapping combined with telecom metadata for AI-targeted cash transfers; identified Africa's 42-language challenge
Nazar Patria (Indonesia): Cited AI-powered TB diagnostics deployed across Indonesia's 17,000 islands; called for balancing protection and innovation
Rafat Hindi (Egypt): Highlighted AI for breast cancer detection and diabetes screening reaching underserved communities
Paula Bogante Zamura (Costa Rica): Challenged the "regulate first" framing for countries investing just 0.3% of GDP in R&D against the OECD average of 2.7%
Omar Alama (UAE): Described the UAE's counterintuitive decision to embrace rather than ban AI tools; advocated continuous proactive assessment over prohibition

Alexandr Wang, Chief AI Officer, Meta

Cited 3.5 billion daily Meta app users (500 million+ in India)
Announced AI-powered Reel translation, WhatsApp Business Agents (set up in 10 minutes), open-sourced omnilingual models covering 1,600+ languages, and collaboration with India's government on the "AI Coach" platform with datasets in 10 major Indian languages
Articulated Meta's "Personal Super Intelligence" vision

Professor Surya Ganguli, Stanford University

Contributed academic perspective bridging AI, neuroscience, and physics to the policy discussions

Policy Announcements & Commitments

Government Commitments

Commitment	Details	Announced By
India's sovereign multilingual AI models launched	Multimodal, multilingual models competitive with frontier models "on the points that matter most"; more than 90% of use cases addressable with smaller, focused models	Ashwini Vaishnaw
Common compute platform as public good	38,000 GPUs at affordable rates for startups, academia, researchers, students; expanding by 20,000 additional GPUs	Ashwini Vaishnaw
New Delhi Frontier AI Impact Commitments	Two voluntary commitments by frontier and Indian AI companies: (1) advancing understanding of real-world AI usage through anonymized/aggregated insights for evidence-based policy, (2) strengthening multilingual and contextual evaluations of AI systems for the Global South	Ashwini Vaishnaw; signed by Anthropic, Google, OpenAI, Microsoft, Sarvam, Bharat Gen AI, Socket, and others
Data residency policy	New policy to attract global data to reside and be processed in India for high-value services — announced during Union Budget	PM Modi / Ashwini Vaishnaw
Nuclear energy reform	Sector recently reformed to provide clean baseload power for AI data centers; over 50% of power generation capacity from clean/renewable sources	PM Modi / Ashwini Vaishnaw
MANAV framework	Ethical AI framework emphasizing human vision as ethical guidance and accountable governance	PM Modi
40-member Independent International Scientific Panel on AI	Leading experts from all regions and disciplines to provide independent scientific advice on AI	António Guterres (UN)
Global Dialogue on AI Governance	First session in Geneva, July 2026; open to all countries, private sector, academia, and civil society	António Guterres (UN)
$3 billion Global Fund for AI	To build basic capacity in developing countries (skills, data, affordable computing, inclusive ecosystems) — described as less than 1% of one major tech company's annual revenue	António Guterres (UN)
Global Network for Exchange and Cooperation on AI Capacity Building	Network focused on the developing world	António Guterres (UN)
France: G7 AI priorities	Inclusive AI vision, children's protection against AI/digital abuse; process to ban social media for children under 15 (India invited to join)	Emmanuel Macron
India-UAE joint AI partnership	Supercomputing cluster, shared data centers, innovation corridor (announced prior week)	Referenced by Macron
Coalition for Sustainable AI	200+ supporters; launching International Challenge for Sustainable AI Models with India and UNESCO	Macron / India
Indo-French AI in Health initiatives	Sorbonne Brain Institute–AIIMS Delhi partnership; Inria–Sanjivini Research Institute (Bangalore)	Macron
Open hardware translation tool for Indian languages and dialects	Joint India-France initiative	Macron
"Curio AI"	Joint India-France launch for sustainable and sovereign AI access	Macron
US AI priorities	Infrastructure, innovation (executive order against onerous regulation), technology sharing with allies; AI Accelerator Partnership with UAE; Pax Silica supply-chain security	Sriram Krishnan (White House)
Africa Forward Summit	Co-hosted by France and Kenya in Nairobi, May 2026; AI and digitalization as key themes	Macron

Program Launches

Program	Description	Lead Organization
Jio Intelligence	Sovereign AI platform with gigawatt-scale data centers (Jamnagar, 120+ MW H2 2026), 10 GW green energy, edge compute, six applications (Shikshak, Arogya, Krishi, Bharat IQ, Frames, Jio-Hotstar AI)	Reliance/Jio
Tata AI-optimized data center with OpenAI	India's first large-scale AI-optimized data center; 100 MW Phase 1, scaling to 1 GW	Tata Group / OpenAI
Tata-AMD AI infrastructure partnership	AMD's AI rack architecture combined with Tata's infrastructure engineering for sustainable high-density AI capacity	Tata Group / AMD
Tata AI data insights platform	Built on diverse Indian datasets atop foundational models	Tata Group
AI Operating System for Industries	Agentic industry solutions for every sector, global launch	TCS / Tata Communications
Tata domain-centric AI chips	AI-optimized semiconductor chips for specific industries; automotive first	Tata Group
Anthropic India office (Bengaluru)	New office with Ireena Vittal as Managing Director; enterprise partnerships with Infosys	Anthropic
Claude Sonnet 4.6	Improved performance across 10 Indic languages	Anthropic
Claude+OpenAgroNet	AI integration for rural farmer advisory	Anthropic
Anthropic Economic Index & Economic Futures Program	Statistical insights on AI's impact on jobs/economy; data sharing with Indian government	Anthropic
Anthropic nonprofit partnerships	With EkStep Foundation, Pratham, Central Square Foundation (education, agriculture, health); with Karya and Collective Intelligence Project (regional language benchmarks)	Anthropic
100 Diffusion Pathways by 2030	Global coalition creating packaged, replicable AI deployment playbooks modeled on India's DPI	Nandan Nilekani / Anthropic / Google / Gates Foundation / UNDP / Kenya
Google AI hub in Visakhapatnam	Full-stack AI hub with gigawatt-scale compute and submarine cable gateway	Google/Alphabet
America India Connect Initiative	Four new submarine fiber optic cable systems between US and India	Google/Alphabet
Google AI Professional Certificate	AI skill-building credential available globally	Google
Horizon 1000	1,000 AI-enabled clinics across Africa	Gates Foundation / OpenAI / Rwanda
Advantage India for AI	Gates Foundation program for AI in India	Gates Foundation
Microsoft Elevate for Educators	Global initiative to equip teachers for AI-integrated education	Microsoft
HCL AI Factory & Physical AI service lines	New AI offerings from HCL Technologies	HCL Technologies
FedEx Import Tool	AI-powered predictive logistics and customs tool, developed first in India for SMEs	FedEx
Meta AI Coach platform	Collaboration with India's government; datasets in 10 major Indian languages	Meta / Government of India
Meta WhatsApp Business Agents	AI-powered business agents, set up in 10 minutes	Meta
Meta omnilingual models	Open-sourced models covering 1,600+ languages	Meta
SynthID	Tool for verifying authenticity of AI-generated content	Google
AI-powered sign language rendering	PM Modi's address simultaneously translated into sign language via AI	Government of India

Funding Announcements

Initiative	Amount	Source
Reliance/Jio sovereign AI infrastructure	~$120 billion (₹10 lakh crore) over 7 years	Reliance Industries / Jio
Adani Group green-energy AI infrastructure	$100 billion (triggering 5 GW, $250 billion ecosystem)	Adani Group
Google India AI infrastructure	~$15 billion	Alphabet / Google
Microsoft Global South AI investment	$50 billion by decade's end	Microsoft
Tata AI data center (Phase 1)	100 MW capacity (scaling to 1 GW)	Tata Group / OpenAI
France data centers (2025)	€58 billion	French government / private sector
European AI commitment	€200 billion	EU-level
Paris AI Action Summit investments	$109 billion (announced; now being delivered)	Multiple (Paris summit signatories)
UN Global Fund for AI (called for)	$3 billion target	UN (proposed)

Policy Frameworks

Framework	Scope	Details
India's Five-Layer AI Stack	National	Applications, models/sovereign layer, compute, infrastructure, energy — comprehensive framework for AI self-sufficiency
MANAV framework	National / Ethical	PM Modi's framework for ethical AI: human vision as guidance, accountable governance, transparency
New Delhi Frontier AI Impact Commitments	International / Voluntary	Two pillars: (1) real-world AI usage insights for evidence-based policy, (2) multilingual/contextual evaluations for Global South effectiveness
UN Independent International Scientific Panel on AI	Global	40 leading experts providing independent scientific advice on AI policy
UN Global Dialogue on AI Governance	Global	Multilateral governance forum; first session Geneva, July 2026
Coalition for Sustainable AI	International	200+ supporters; International Challenge for Sustainable AI Models with India, France, and UNESCO
France G7 AI priorities	G7 / International	Inclusive AI vision; children's protection; social media restrictions for under-15s
India three-pillar sovereignty model (Adani)	National	Energy, compute/cloud, services as pillars of AI sovereignty
US three-pillar AI approach	US / Allied	Infrastructure, innovation (anti-regulation executive order), technology sharing with allies
AI Accelerator Partnership	US-UAE	Joint infrastructure and innovation partnership
Pax Silica	US / Allied	Supply-chain security initiative for AI hardware/chips

Key Topics Covered

AI sovereignty and national AI strategy — compute as public good, sovereign models, data residency, avoiding structural dependency on foreign AI platforms
Investment in AI infrastructure — gigawatt-scale data centers, GPU compute platforms, submarine cables, edge computing, semiconductor manufacturing
Multilingual and inclusive AI — performance across Indian regional languages, African languages, and Global South contexts; Small Language Models versus Large Language Models
AI for development — healthcare (diagnostics, clinical documentation, drug discovery), education (tutors, skill-building), agriculture (weather forecasts, pest detection, rural advisory), financial inclusion
Digital Public Infrastructure (DPI) — India Stack (Aadhaar, UPI, Bhashini), health IDs, agri-stacks, replication in 40+ countries
AGI/superintelligence timelines and feasibility — sharply contested among summit speakers
Open-source versus proprietary AI models — geopolitical and economic implications, safety trade-offs
AI governance and international cooperation — IAEA-like bodies, UN scientific panels, multilateral dialogues, voluntary commitments, executive orders
AI and jobs — displacement versus augmentation, evidence-based analysis, workforce reskilling, talent pipeline development
Energy and sustainability for AI — clean/renewable energy, nuclear power, data center energy demands, water consumption, sustainable AI models
Edge AI and frugal computing — running LLMs on low-power devices, inference cost as primary barrier
Trust, safety, and responsible AI — content authenticity/provenance, bias, children's protection, cybersecurity, human oversight
AI in the physical world — autonomous vehicles, smart connected devices, car-to-car communication, industrial IoT
Semiconductors and chip design — domain-specific AI chips, India's SemCon India policy, NXP's Kinara acquisition
Agentic AI and enterprise transformation — workflow automation, enterprise AI operating systems, non-human identity management
AI and scientific discovery — AlphaFold, protein structure prediction, drug development, climate modeling
Children's safety and AI — social media bans for minors, unregulated AI exposure, digital abuse prevention
Quantum computing — French investments in Pasqal, Quandela, Alice & Bob, Quobly
Global inequality and the digital/AI divide — avoiding repeat of electricity's 144-year exclusion, funding gaps for SDGs, R&D investment disparities

Speakers & Organizations Mentioned

Speakers / Named Individuals

Name	Title / Affiliation
Narendra Modi	Prime Minister of India
Ashwini Vaishnaw	Minister for Electronics & Information Technology, Government of India
Mukesh Dhirubhai Ambani	Chairman & Managing Director, Reliance Industries / Jio
N. Chandrasekaran	Chairman, Tata Group
Sam Altman	CEO, OpenAI
Sundar Pichai	CEO, Alphabet and Google
Dario Amodei	CEO, Anthropic
Sir Demis Hassabis	Co-founder & CEO, Google DeepMind; Nobel Laureate
Yann LeCun	Executive Chairman, Advanced Machine Intelligence Labs (AMI Labs); "Godfather of Deep Learning"
António Guterres	Secretary-General, United Nations
Emmanuel Macron	President of France
Nandan Nilekani	Co-founder, Infosys; Architect of Aadhaar
Rishi Sunak	Former Prime Minister, United Kingdom
Brad Smith	Vice Chair and President, Microsoft
Vinod Khosla	Founder, Khosla Ventures
Jeet Adani	Director, Adani Digital Labs
Roshni Nadar Malhotra	Chairperson, HCL Technologies
Shantanu Narayen	CEO, Adobe
Rishad Premji	Executive Chairman, Wipro
Julie Sweet	Chair and CEO, Accenture
Rajesh Subramanyan	CEO, FedEx
Nikesh Arora	CEO, Palo Alto Networks
Arthur Mensch	CEO, Mistral AI
Ankur Vora	Gates Foundation representative
Vishal Sikka	Founder & CEO, VNI
Roy Jakobs	CEO, Philips
Lars Reger	CTO, NXP Semiconductors
Amit Zavery	President & CPO, ServiceNow
Martin Schroeter	CEO, Kyndryl
Sunil Mittal	Founder/Chairman, Bharti Airtel
Bejul Somaia	Managing Director, Lightspeed Venture Partners
Olivier Blum	CEO, Schneider Electric
Sriram Krishnan	Senior Policy Adviser for AI, White House
Cina Lawson	Minister, Togo
Nazar Patria	Minister/Official, Indonesia
Rafat Hindi	Minister/Official, Egypt
Paula Bogante Zamura	Minister/Official, Costa Rica
Omar Alama	Minister/Official, UAE
Ireena Vittal (Ireena Ghosh)	Managing Director, Anthropic India
Professor Surya Ganguli	Stanford University
Alexandr Wang	Chief AI Officer, Meta

Organizations, Companies & Institutions

Technology Companies: Alphabet/Google, Google DeepMind, OpenAI, Anthropic, Meta, Microsoft, Mistral AI, AMD, NXP Semiconductors, Adobe, Palo Alto Networks, ServiceNow, Kyndryl, Schneider Electric, Philips, Hugging Face, Poolside, VNI

Indian Conglomerates & Companies: Reliance Industries/Jio, Tata Group/TCS/Tata Communications, Adani Group/Adani Digital Labs, Infosys, HCL Technologies, Wipro, Bharti Airtel, Sarvam, Bharat Gen AI, Socket, Kinara, Wadhwani AI

Startups & Emerging Companies: Mistral AI, Pasqal, Quandela, Alice & Bob, Quobly, Arago/Armaton AI, CK12.org

Venture Capital / Investment: Khosla Ventures, Lightspeed Venture Partners

Logistics / Industry: FedEx, Accenture, Amul

International Organizations: United Nations, UNESCO, UNDP

Foundations & Nonprofits: Gates Foundation, EkStep Foundation, Pratham, Central Square Foundation, Karya, Collective Intelligence Project, Azim Premji Foundation

Government Bodies: Government of India (MeitY), White House (US), Government of France, Government of Togo, Government of Indonesia, Government of Egypt, Government of Costa Rica, Government of UAE, Government of Rwanda, Government of Kenya, Government of El Salvador, Government of Uganda, Government of Ghana, BRICS, G7, EU

Academic & Research Institutions: Stanford University, Sorbonne Brain Institute, AIIMS Delhi, Inria, Sanjivini Research Institute (Bangalore), AMI Labs (Yann LeCun)

Infrastructure / Platforms: India Stack, Aadhaar, UPI, Bhashini, Unified Lending Interface (ULI), Bangalore Airport (self-healing AI operations)

Countries Referenced

India, France, United States, United Kingdom, United Arab Emirates, China, Togo, Indonesia, Egypt, Costa Rica, Rwanda, Kenya, El Salvador, Uganda, Ghana, Thailand, Malaysia, Sweden, Netherlands, Greece, Ethiopia

Technical Concepts & Resources

AI Models & Systems

Claude Sonnet 4.6 — Anthropic's model with improved 10 Indic language performance
Claude+OpenAgroNet — AI integration for rural agricultural advisory
AlphaFold — Google DeepMind's protein structure prediction system (Nobel Prize-winning)
Neural GCM — Google's AI weather/climate forecasting model
Gemini — Google DeepMind's foundation models (Reliance Jio partnership)
ChatGPT / Codex — OpenAI's consumer and developer AI products
SynthID — Google's content authenticity verification tool
DeepSeek — Chinese AI model referenced by Macron
Sovereign multilingual models — India's government-launched models at the summit
Small Language Models (SLMs) — Task-specific, smartphone-capable models (India's strategy)
Meta omnilingual models — Open-sourced models covering 1,600+ languages

Frameworks & Concepts

Five-Layer AI Stack — India's national AI architecture: applications, models, compute, infrastructure, energy
MANAV framework — PM Modi's ethical AI framework
New Delhi Frontier AI Impact Commitments — Two-pillar voluntary commitments on usage insights and multilingual evaluations
"Country of geniuses in a data center" — Dario Amodei's metaphor for near-future AI capability
Moore's Law for intelligence — Dario Amodei's characterization of exponential AI capability growth
Moravec's Paradox — Referenced by Yann LeCun; easy physical tasks are hard for AI, hard cognitive tasks are easy
"Jagged frontier" — Vishal Sikka's concept of non-uniform AI effectiveness across tasks
Digital Public Infrastructure (DPI) — India Stack model: Aadhaar, UPI, Bhashini, health IDs
India Stack — Aadhaar (1.4B digital identities), UPI (20B monthly transactions), digital health IDs (500M)
Agentic AI — Autonomous AI agents performing complex multi-step workflows
Edge AI — AI inference on local devices rather than cloud; LLMs on low-power hardware
Frugal inference — Cost-efficient AI computation for mass deployment
Non-human identity management — Security framework for AI agents and automated systems
Content provenance / watermarking — Authenticity verification for AI-generated content
Personal Super Intelligence — Meta's vision for individualized AI assistants

Platforms & Initiatives

Jio Intelligence — Reliance's sovereign AI platform (six applications: Shikshak, Arogya, Krishi, Bharat IQ, Frames, Jio-Hotstar AI)
100 Diffusion Pathways by 2030 — Global AI deployment replication coalition
Horizon 1000 — Gates Foundation's 1,000 AI-enabled clinics initiative for Africa
Advantage India for AI — Gates Foundation program for India
Microsoft Elevate for Educators — Global AI education initiative
Google AI Professional Certificate — AI skills credential
Curio AI — India-France sustainable and sovereign AI access platform
AI Coach — Meta/India government collaboration platform (10 Indian languages)
FedEx Import Tool — AI-powered predictive logistics for SMEs
Anthropic Economic Index / Economic Futures Program — AI-and-jobs statistical insights
International Challenge for Sustainable AI Models — India/France/UNESCO initiative
Coalition for Sustainable AI — 200+ supporter international coalition
America India Connect Initiative — Four new US-India submarine cable systems
AI Accelerator Partnership — US-UAE joint AI initiative
Pax Silica — US supply-chain security initiative for AI hardware
Stargate — US mega-scale AI infrastructure project (referenced by Macron)
Unified Lending Interface (ULI) — Indian financial inclusion platform
LUMI/Alisoc — European exascale supercomputer
SemCon India — India's semiconductor policy framework
Bhashini — India's multilingual translation platform
CK12.org — AI-powered educational platform (400M students, CBSE/NEP-compatible)

Infrastructure Metrics & Benchmarks

38,000 GPUs (expanding by 20,000) — India's common compute platform
100 MW → 1 GW — Tata/OpenAI data center scaling roadmap
120+ MW — Jio Jamnagar data center (operational H2 2026)
10 GW — Jio green energy surplus
5 GW — Adani integrated energy-compute ecosystem
10-billion-parameter LLM at 7 watts — NXP edge AI demonstration
50 billion — Predicted smart connected devices within a decade
2 petabytes/day — FedEx data generation
55x growth — ServiceNow agentic workflow adoption after security controls
250x productivity improvement — Vishal Sikka's generative AI coding demonstration
Stanford AI Index — Benchmark cited for India's AI leadership ranking