🗞️ Why in News The India AI Impact Summit 2026 (February 16–21, Bharat Mandapam, New Delhi), the fourth global AI summit after Bletchley Park (2023), Seoul (2024), and Paris (2025), saw PM Modi unveil the MANAV framework for human-centric AI. Meanwhile, NITI Aayog warned that AI could wipe out 2.7 million tech jobs by 2031, and India’s plan to become a global data centre hub raises serious concerns about water, energy, and carbon emissions.

India AI Impact Summit 2026

  • Venue: Bharat Mandapam, New Delhi
  • Dates: February 16–21, 2026
  • Significance: First AI summit in the series hosted by a Global South nation
  • Previous summits: Bletchley Park AI Safety Summit (UK, 2023) → AI Seoul Summit (South Korea, 2024) → AI Action Summit (Paris, 2025; co-chaired by PM Modi and President Macron)
  • Key outcome: New Delhi Declaration on AI governance
  • Bill Gates pulled out as keynote speaker

The MANAV Framework

PM Modi unveiled MANAV (meaning “human” in Hindi), India’s vision for AI governance:

Letter Pillar Meaning
M Moral and Ethical Systems AI must be rooted in fairness, transparency, and human oversight
A Accountable Governance Transparent rules and robust oversight mechanisms
N National Sovereignty Data sovereignty, algorithmic sovereignty, and digital infrastructure sovereignty
A Accessible and Inclusive AI must not be a monopoly but a multiplier, especially for the Global South
V Valid and Legitimate Trust, safety, legality, AI systems must be verifiable, lawful, and transparent (especially regarding deepfakes and synthetic media)
Key Principles
  • Transform AI from machine-centric to human-centric
  • AI must not reduce human beings to “mere data points or raw material”
  • Democratise AI as a medium of inclusion and empowerment
  • Ensure “the welfare of all, the happiness of all” (Sarve Bhavantu Sukhinah philosophy)

⊕ Supplementary UPSC Concepts (Additional Material)

The following sections provide supplementary explanatory material with key concepts for UPSC beyond the article.

UN “AI for Good” Initiative: Context for MANAV

  • Launched: 2017 by ITU (International Telecommunication Union), the UN’s specialised agency for digital technologies
  • Partners: 53+ UN organisations, co-convened with Government of Switzerland
  • Goal: Identify practical AI applications to advance the UN Sustainable Development Goals (SDGs)
  • Scale: 37,000 contributors from 180+ countries
  • Annual summit: AI for Good Global Summit held in Geneva since 2017
  • Relevance: MANAV framework draws on similar principles but frames them as an Indian philosophical contribution, the article notes “not many here appear to be aware” of the UN’s existing initiative

IndiaAI Mission: India’s AI Infrastructure Push

  • Launched: March 2024
  • Budget: ₹10,371.92 crore (~$1.14 billion) over 5 years
  • Compute capacity: ₹4,563.36 crore allocated (44% of total budget, largest single component)
  • GPUs deployed: 38,000 (target was initially 10,000); 70% high-end NVIDIA H100, 30% older models
  • Subsidised rate: ₹65/hour for AI startups and researchers
  • Budget trajectory: FY 2024-25: ₹551.75 crore → FY 2025-26: ₹2,000 crore (1,056% increase in revised estimates)
  • 7 pillars: Compute infrastructure, AI innovation centres, datasets platform, application development, skilling, startup financing, responsible AI

Sarvam AI: India’s Sovereign LLM Effort

  • Selected by MeitY (April 2025) under IndiaAI Mission to develop an indigenous foundational AI model
  • Sarvam-30B: 30-billion parameter model (mixture-of-experts design), released February 18, 2026
  • Sarvam-105B (“Indus”): 105-billion parameter open-source LLM, released February 20, 2026
    • Trained from scratch under IndiaAI Mission
    • Supports all 22 official Indian languages (Eighth Schedule)
    • Hosted on Hugging Face under Apache License
  • The article notes Sarvam AI “failed to evoke any excitement” at the summit, reflecting the gap between announcements and ground-level AI capability

What is a Foundational/Large Language Model (LLM)?

  • LLM: A deep learning model trained on massive text datasets to understand and generate human language
  • “Foundational” means it serves as a base model that can be fine-tuned for specific applications
  • Parameters: The “weights” that the model learns during training, more parameters generally means more capability (GPT-4: ~1.8 trillion; Sarvam-105B: 105 billion)
  • Sovereign AI: The concept that nations should develop their own AI models and infrastructure to avoid dependence on foreign (primarily US) companies, a key concern for data privacy and national security

The Jobs Crisis: NITI Aayog Warning

The Report
  • Title: “Roadmap for Job Creation in the AI Economy”
  • Released: October 2025
  • Developed by: NITI Aayog’s Frontier Tech Hub with NASSCOM and BCG
Key Findings
  • AI could wipe out 2.7 million jobs in tech services by 2031
  • Tech services employ 13% of total workforce and over 30% of white-collar talent
  • Headcount could fall from 7.5–8 million to 6 million by 2031
  • Customer services sector: 2–2.5 million to 1.8 million
  • Centre of Advanced Study in India: Over 60% of formal sector jobs vulnerable to AI automation by 2030
  • India’s share of granted AI patents fell from 8–10% (2010) to under 5% (2023)
The Three Major Challenges
  1. Scale of jobs at risk: millions in the flagship tech sector
  2. Fundamental shortcomings in education and skills: curriculum and skilling cycles too slow
  3. Growing shortage of AI talent: paradox for a country with the world’s largest pool of young digital talent
Proposed Solution
  • India AI Talent Mission: nationally coordinated programme to equip the workforce for AI disruption
  • AI is “advancing faster than policy, curriculum, and skilling cycles can adapt to”
  • Risk: not just irreversible job losses, but “broader societal disruption, economic marginalisation, and weakening of global competitiveness”
  • The report also notes potential to create 4 million new jobs: but only with urgent reskilling

The Data Centre Gamble: Environmental Cost

The Plan
  • India pitching itself as a global hub for AI data centres
  • PM invited “all of world’s data to reside in India”
  • Expected investment: $200 billion: Google, Amazon, Microsoft have made commitments
  • Google: Setting up a 1 GW data centre in Visakhapatnam (243 hectares)
Why It Won’t Solve the AI Problem
  • AI behemoths do not share data or technology: hosting their data centres doesn’t transfer AI capability to India
  • Data centres are irrelevant to developing foundational LLMs: India needs its own models, not server farms for US companies
  • Employment generation will be minimal (data centres are largely automated)
The Environmental Cost
Resource Current (2024–25) Projected (2030)
Data centre capacity ~1.5 GW 4.5–17 GW
Electricity consumption ~0.5% of national grid 8% of national energy
Water consumption 150 billion litres/year 358 billion litres/year
  • A 100 MW data centre consumes over 2 million litres of water daily
  • 60–80% of India’s data centres will face high water stress (S&P Global)
  • Most centres in water-scarce cities: Mumbai, Chennai, Hyderabad, Bengaluru
  • India still relies on coal-based power: data centres add to carbon emissions
  • More than half of India’s districts face high risk from extreme heat → data centres in these areas need higher cooling loads
  • No carbon emissions framework specific to data centres exists in India
  • No mandatory EIA for data centres

Critical Evaluation for UPSC Mains

The Fundamental Contradiction
  • India simultaneously claims AI leadership (MANAV, sovereign AI) while lacking the basic infrastructure: specialised chips (GPUs are all imported), trained talent, and foundational models
  • The data centre strategy serves foreign AI companies, not Indian AI sovereignty
  • MANAV’s “national sovereignty” pillar is undermined by the data centre plan, India becomes a service provider (hosting data) rather than an AI power (creating models)
  • The jobs crisis timeline (5 years) is alarmingly short, reskilling 2.7 million workers requires institutional speed that India’s education system has not demonstrated
The Digital Divide Dimension
  • AI disruption will disproportionately affect lower-skilled tech workers: the very population that the IT boom lifted into the middle class
  • Rural India remains disconnected, 59% internet penetration (2024), significant urban-rural gap
  • MANAV’s “accessible and inclusive” pillar requires bridging this gap first
The Environmental Trade-off
  • Data centres consuming 358 billion litres by 2030 while India faces worsening water stress (Jal Jeevan Mission struggling to deliver)
  • Coal-powered data centres contradicting Net Zero 2070 target
  • The irony: hosting “green” AI infrastructure using coal-generated electricity

UPSC Angle

  • Prelims: MANAV framework (full form), India AI Impact Summit 2026 (venue, dates), IndiaAI Mission (budget: ₹10,372 crore), NITI Aayog AI jobs report, Sarvam AI, LLM, AI for Good (ITU, 2017), Bletchley Park/Seoul/Paris AI summits, NASSCOM, data centre water consumption, sovereign AI
  • Mains GS-2: Governance, AI governance frameworks, MANAV vs international frameworks (EU AI Act, UNESCO AI Ethics), digital sovereignty, government initiatives (IndiaAI Mission)
  • Mains GS-3: Economy, AI disruption of tech services, job losses vs job creation, India’s AI patent decline, data centre economics; Environment, data centre water/energy footprint, coal-based power for AI, carbon emissions; Science & Technology, LLMs, foundational models, GPU infrastructure, AI sovereignty
  • Essay: “India’s AI future cannot be built on other nations’ data, it must be built on its own talent and values”

📌 Facts Corner: Knowledgepedia

India AI Impact Summit 2026:

  • Dates: February 16–21, 2026; Venue: Bharat Mandapam, New Delhi
  • 4th global AI summit (after Bletchley Park 2023, Seoul 2024, Paris 2025)
  • First AI summit hosted by a Global South nation
  • New Delhi Declaration on AI governance adopted

MANAV Framework:

  • M: Moral and Ethical Systems
  • A: Accountable Governance
  • N: National Sovereignty
  • A: Accessible and Inclusive
  • V: Valid and Legitimate
  • Unveiled by PM Modi at the summit

NITI Aayog AI Jobs Report (October 2025):

  • Title: “Roadmap for Job Creation in the AI Economy”
  • Developed with NASSCOM and BCG
  • 2.7 million tech jobs at risk by 2031
  • Tech services: 7.5–8 million → 6 million (by 2031)
  • 60% of formal sector jobs vulnerable to AI automation by 2030
  • India’s AI patents: 8–10% (2010) → under 5% (2023)
  • Solution proposed: India AI Talent Mission

IndiaAI Mission:

  • Launched: March 2024; Budget: ₹10,371.92 crore over 5 years
  • Compute: ₹4,563.36 crore (44% of budget)
  • GPUs: 38,000 deployed; 70% NVIDIA H100
  • Subsidised rate: ₹65/hour

Sarvam AI:

  • Selected by MeitY under IndiaAI Mission (April 2025)
  • Sarvam-30B: 30B parameters (Feb 18, 2026)
  • Sarvam-105B “Indus”: 105B parameters, 22 Indian languages (Feb 20, 2026)
  • Open-source, Apache License, on Hugging Face

Data Centre Environmental Impact:

  • Current capacity: ~1.5 GW; projected by 2030: 4.5–17 GW
  • Water: 150 billion litres (2024) → 358 billion litres (2030)
  • 100 MW centre = 2 million litres water/day
  • 60–80% of India’s data centres face high water stress
  • Google: 1 GW centre in Visakhapatnam (243 ha)
  • No EIA or carbon framework for data centres in India

AI for Good (UN/ITU):

  • Launched: 2017 by ITU (International Telecommunication Union)
  • 53+ UN organisations as partners
  • 37,000 contributors from 180+ countries
  • Annual summit in Geneva since 2017

Other Relevant Facts:

  • EU AI Act: world’s first comprehensive AI regulation (entered force August 2024)
  • UNESCO Recommendation on Ethics of AI: adopted November 2021 (193 countries)
  • Global AI summit series: Bletchley Park (2023) → Seoul (2024) → Paris (2025) → New Delhi (2026)
  • India internet penetration: ~59% (2024); urban-rural digital divide persists
  • Net Zero 2070: India’s long-term climate target
  • NASSCOM: National Association of Software and Services Companies (India’s IT industry body)

Sources: Down to Earth, NITI Aayog, PIB, IndiaAI Mission, ITU