The India AI Impact Summit wrapped its fourth day today with a staggering tally: over $200 billion in combined investment commitments, voluntary frameworks signed by every major AI lab, and 20 heads of state participating in what organizers called the first major global AI summit hosted in the Global South.
The numbers are impressive. The questions they raise are harder.
The Commitment Avalanche
The investment pledges came fast:
- Adani Group: $100 billion by 2035 for renewable-powered AI data centers, potentially catalyzing a $250 billion ecosystem
- Microsoft: $50 billion for Global South AI infrastructure by decade’s end
- Google: $15 billion for Indian AI infrastructure, plus the America-India Connect subsea cable project
- Yotta Data Services: $2 billion for 20,736 Nvidia Blackwell Ultra GPUs, creating one of Asia’s largest AI computing clusters
- Blackstone: $600 million in Indian cloud startup Neysa
- Indian Government: $1.1 billion earmarked for an AI venture capital fund
OpenAI announced a strategic partnership with Tata Group, becoming the first customer of TCS’s HyperVault data center business with 100 megawatts of initial capacity scaling to 1 gigawatt. Anthropic opened a Bengaluru office, its second in Asia. Google DeepMind partnered with India’s Anusandhan National Research Foundation to give researchers access to tools like AlphaGenome and AI Co-scientist.
Union IT Minister Ashwini Vaishnaw announced India is on track to triple its GPU capacity to 100,000 by year’s end.
What the “New Delhi Frontier AI Commitments” Actually Say
The summit’s headline policy output is the New Delhi Frontier AI Commitments - a voluntary framework signed by OpenAI, Google, Anthropic, and Indian companies including Sarvam, Bharatjan, Yani, and Soket.
The commitments have two focus areas:
First, using anonymized and aggregated data to support evidence-based policymaking on jobs and skills. This sounds reasonable until you consider that “anonymized and aggregated” is doing a lot of work - the same framing that enabled Cambridge Analytica.
Second, strengthening multilingual and use-case evaluations to ensure AI works across languages and cultures. This addresses a real gap. Current AI benchmarks are overwhelmingly English-centric, and models perform significantly worse in Hindi, Tamil, or Swahili than in English.
But both commitments are voluntary. There’s no enforcement mechanism, no penalties for non-compliance, and no independent verification. The companies signing them retain complete discretion over implementation.
India’s Balancing Act
Prime Minister Modi framed the summit around placing “the aspirations and priorities of the Global South at the centre of AI governance.” French President Macron called India “a safe space for innovation” and praised its ability to build public digital infrastructure “no other country could.”
The rhetoric positions India as a bridge between Western AI development and Global South needs. That’s a valuable role - but it requires navigating contradictions.
India wants Western investment and technology transfer while developing “Made in India” ethical guidelines rather than importing Western standards. It wants to compete with China while remaining entangled in Chinese supply chains. It wants inclusive AI development while its government considers regulations that could concentrate power in established players.
The summit itself illustrated these tensions. A university was expelled from the exhibition after presenting a Chinese-made Unitree robot dog as an Indian innovation called “Orion.” The incident became a political flashpoint, with opposition leader Rahul Gandhi calling the summit a “disorganized PR spectacle” where “Chinese products were being showcased while India’s own talent and data potential were ignored.”
The episode revealed an uncomfortable truth: India’s AI ambitions remain dependent on hardware ecosystems that China and the US dominate.
What “Global South” Means Here
Representatives from Uganda, Indonesia, Ghana, and other developing nations attended sessions on democratizing AI compute and digital infrastructure. The stated goal: shared learning, regulatory harmony, and infrastructure ties for economic and social progress.
But there’s a vast gap between India and most Global South nations. India has 100 million weekly ChatGPT users - the second-largest base globally. It has engineering talent, domestic capital, and increasingly, domestic compute. Countries like Uganda are starting from near-zero AI infrastructure.
The question is whether summit outcomes benefit that broader Global South or primarily benefit India’s positioning as a regional tech hub. Microsoft’s $50 billion pledge covers “the Global South” - but the bulk of that investment will flow through existing hubs like India, Mexico, and select African and Southeast Asian cities with adequate infrastructure.
For countries with limited compute power, the summit offered collaboration opportunities but few concrete resource commitments. The path from “shared learning” to actual GPU clusters remains unclear.
The AI Lab Positioning
Every major AI lab showed up with announcements:
Sam Altman called India a potential “full-stack AI leader” and announced “OpenAI for India” with offices opening in Mumbai and Bengaluru. India has emerged as OpenAI’s second-largest user base globally, making the engagement commercially sensible.
Dario Amodei said India has an “absolutely central role” in shaping AI’s future and predicted 25% economic growth potential from AI adoption. Anthropic positioned its expansion as spreading “AI’s benefits across the Global South, starting with India.”
Sundar Pichai announced a suite of initiatives including the subsea cable project, partnerships with the Indian government on AI training for 20 million public servants, and $60 million in research challenges.
These are real investments. But the framing matters. The AI labs aren’t philanthropists - they’re commercial enterprises securing access to a massive, fast-growing market. “Global South” becomes a branding opportunity as much as a development priority.
What Wasn’t Addressed
Several issues received minimal attention:
Data sovereignty: Indian users’ data flowing to US-based AI systems. The “anonymized and aggregated” language in the Frontier AI Commitments doesn’t address where data is processed or stored.
Concentration risk: The investments flow overwhelmingly to a few large players - Adani, Tata, Yotta. The summit celebrated startup culture while potentially entrenching oligopolies.
Labor displacement: Altman predicted “new and better jobs” from AI. The more pressing question for India’s hundreds of millions of service-sector workers is which jobs disappear and how fast.
Environmental impact: Adani’s $100 billion pledge is for “renewable-powered” data centers, but the Khavda solar project providing that power is itself controversial. Data centers consume enormous amounts of water for cooling in a water-stressed country.
Governance enforcement: Voluntary commitments with no verification mechanism have a poor track record across industries.
The Real Test
The summit succeeded at what summits do: convening powerful actors, generating headlines, and producing aspirational documents. India positioned itself as the Global South’s voice in AI governance, and Western tech companies secured access to a market they desperately want.
Whether anything substantive emerges depends on what happens next. The Leaders’ Declaration coming at the summit’s close will reveal whether countries agreed to concrete actions or just principles. The voluntary commitments only matter if companies actually implement them - and someone tracks whether they do.
Most importantly, the infrastructure investments need to materialize. A $100 billion pledge by 2035 sounds transformative. But Adani’s commitments have faced scrutiny before, and decade-long timelines conveniently exceed electoral cycles and executive tenures.
The Bottom Line
The India AI Impact Summit demonstrated the Global South’s growing importance in AI development - and the limits of that importance when Western companies and Indian conglomerates control the resources.
For India specifically, the summit was a success: international validation, massive investment pledges, and positioning as a bridge between development paradigms. For the broader Global South, the benefits remain aspirational. And for anyone skeptical of voluntary corporate commitments backed by no enforcement mechanism, the New Delhi Frontier AI Commitments look like another round of declarations that sound better than they bind.
The infrastructure is coming. The question is who controls it, who benefits from it, and whether the governance frameworks being celebrated this week have any teeth at all.