deep-dive February 21, 2026

The $400 Billion Bet: Inside India’s Moment of Maximum AI Ambition

The India AI Impact Summit just closed with $250 billion in infrastructure pledges, a Delhi Declaration signed by 70+ nations, and every major AI CEO on a stage in New Delhi. The numbers are staggering. The question is whether India can actually build what it has promised - and why the answer matters far beyond its borders.

There is a phrase that Indian Prime Minister Narendra Modi coined for the New Delhi AI Impact Summit: MANAV AI - “of the humans, by the humans, for the humans.” It is a deliberate riff on Lincoln, pitched at a gathering that included Sam Altman, Sundar Pichai, Dario Amodei, Demis Hassabis, and every other AI executive who matters. But beneath the rhetorical flourish lies something more substantive than the usual summit optimism. When the closing gavel fell on Saturday afternoon in Bharat Mandapam, India had accumulated $250 billion in infrastructure investment commitments, $150 billion in deep-tech venture capital pledges, a geopolitical declaration signed by more countries than any prior AI summit in history, and - crucially - a formal trade architecture with the United States that could reshape the global technology supply chain.

This was not a conference. It was a reconfiguration of where the AI industry’s center of gravity sits.

The Money: A Number That Demands Context

The headline figure - $250 billion in infrastructure pledges - is almost too large to parse without disaggregating it. The largest single commitment came from Reliance Industries and its telecom arm Jio: $109.8 billion over seven years to build AI and data infrastructure, announced by billionaire chairman Mukesh Ambani. Adani Group followed with a $100 billion commitment for renewable energy-powered AI data centers by 2035, with the company projecting that its investment would catalyze an additional $150 billion across server manufacturing, sovereign cloud platforms, and related industries - creating what it estimated as a $250 billion AI infrastructure ecosystem in India within a decade.

From the American side, Microsoft disclosed it was on track to invest $50 billion across the Global South by the end of the decade, building on the $17.5 billion in India-specific investments it announced last year. OpenAI and AMD both announced partnerships with Tata Group to build AI capabilities; Tata Consultancy Services separately signed OpenAI as the first customer for its data center unit under the Stargate initiative. Larsen & Toubro - India’s massive infrastructure conglomerate - announced a proposed joint venture with Nvidia to build AI-ready data center infrastructure at scale. Indian firm Yotta Data Services committed more than $2 billion to build one of Asia’s largest AI computing hubs powered by Nvidia’s latest Blackwell Ultra chips. Blackstone participated in a $600 million equity raise for Indian AI infrastructure firm Neysa.

Nvidia itself announced expanded partnerships with Indian venture capital firms, deepening its exposure to the startup ecosystem. The company’s presence was strategic, not ceremonial: Nvidia recognizes that the next phase of its growth will come from markets that are building AI infrastructure from scratch, not from replacing existing architecture.

The aggregate is staggering: more than $400 billion in direct and catalyzed capital commitments from a four-day summit. For context, the entire U.S. CHIPS Act allocated $52 billion.

The Declaration: More Than Theater

Alongside the money came the Delhi Declaration, a non-binding but politically significant framework for how AI should be developed and governed globally. By Saturday, 70 nations had signed it; the government expected the number to cross 80 before the final document was released.

The scale of signatories matters because it marks a break from recent history. The Bletchley Park Summit in 2023 attracted 30 signatories. Seoul in 2024 had 11. Paris in 2025 - which India co-hosted - managed 58, but notably without the United States or United Kingdom. Washington declined to sign the Paris Declaration, citing concerns over excessive regulation. London objected that the document lacked sufficient clarity on national security risks.

India, it appears, threaded a needle that France could not. The New Delhi Declaration attracted the United States as a signatory - a diplomatic achievement that IT Minister Ashwini Vaishnaw described with careful understatement as including “all major countries, people who matter in AI.” The shift reflects both India’s positioning as a neutral interlocutor in an era of US-China tech bifurcation, and the Trump administration’s willingness to engage multilaterally on AI governance when the framework does not impose regulatory constraints it finds objectionable.

The declaration’s significance is political, not legal. It does not bind signatories to specific policies. What it does is establish India as a convening authority in global AI governance - a role that carries real weight in how standards, interoperability norms, and regulatory templates propagate through the 70-plus nations that signed.

The Architecture: Pax Silica and the Trade Dimension

The most consequential structural development at the summit received less coverage than the investment headlines but may prove more durable. Representatives of the U.S. and Indian governments signed the Pax Silica agreement - a U.S.-led initiative launched by the Trump administration aimed at securing the global supply chain for silicon-based technologies. The signing came alongside a framework for an interim US-India trade deal that would lower tariffs and increase economic cooperation between the two countries.

Read together, these agreements sketch the architecture of a new alignment. The U.S. is actively constructing a technology supply chain that excludes China and includes trusted partners - India chief among them. India’s combination of engineering talent, domestic market scale, and democratic governance makes it uniquely attractive as a Pax Silica anchor. The $18 billion in chip manufacturing projects that India has already approved are the industrial foundation that makes this alignment credible rather than aspirational.

Microsoft President Brad Smith put the strategic logic plainly in an interview at the summit: “If you look at the engineering talent, you quickly conclude India too can be a place where models are developed.” He predicted that “a variety of different DeepSeek moments” would come from India in the future - a reference to the Chinese AI lab’s surprise competitive breakthrough earlier this year. The comparison is deliberate. DeepSeek demonstrated that frontier AI development need not be concentrated in Silicon Valley. Smith is suggesting India could produce its own version of that disruption.

The Skeptics Are Not Wrong

The gap between summit promises and ground-level reality in India is wide enough to be honest about. Udith Sikand, a senior emerging markets analyst, offered a corrective at the margins of the event: “India is making splashy attempts to kickstart its belated AI push, but it is doing so primarily by offering headline-grabbing sops without addressing many of the underlying difficulties of actually doing business in India.” The observation is not uncharitable. India’s regulatory environment, logistics infrastructure, power grid reliability, and venture capital ecosystem all lag the ambitions being announced from the stage at Bharat Mandapam.

Anirudh Suri, founding partner of the India Internet Fund, identified the specific gap that matters most: “What we’ve not maybe seen as much of right now is venture capital and private equity money to come in to invest in Indian entrepreneurs in the AI space.” The infrastructure commitments from Reliance and Adani build data centers. The Microsoft and Nvidia partnerships wire those centers with compute. But the layer that converts compute into competitive AI companies - patient venture capital willing to fund Indian AI founders through long development cycles - remains underdeveloped.

The controversy that surrounded the summit offered a more pointed illustration of the gap between aspiration and execution. Bill Gates withdrew from the event after public backlash over his past relationship with Jeffrey Epstein - a reminder that India’s tech diplomacy operates in the same information environment as everyone else, subject to the same reputational dynamics. An Indian university was criticized after falsely claiming it had invented a commercially available Chinese-made robot dog. The story was minor in itself but symptomatic of a broader pressure to demonstrate AI capability that India does not yet possess at scale.

Why This Moment Is Different

Skepticism about India’s technology ambitions has a long and often well-founded history. The question is whether this summit represents another cycle of announcement followed by underdelivery, or whether something structurally different is happening.

See also: The $110 Billion Bet: OpenAI.

For related context, see Nvidia’s Plan to Make Every Cell Tower an AI Supercomputer.

The scale of signatories matters because it marks a break from recent history. The Bletchley Park Summit in 2023 attracted 30 signatories. Seoul in 2024 had 11. Paris in 2025 - which India co-hosted - managed 58, but notably without the United States or United Kingdom. Washington declined to sign the Paris Declaration, citing concerns over excessive regulation. London objected that the document lacked sufficient clarity on national security risks.

India, it appears, threaded a needle that France could not. The New Delhi Declaration attracted the United States as a signatory - a diplomatic achievement that IT Minister Ashwini Vaishnaw described with careful understatement as including “all major countries, people who matter in AI.” The shift reflects both India’s positioning as a neutral interlocutor in an era of US-China tech bifurcation, and the Trump administration’s willingness to engage multilaterally on AI governance when the framework does not impose regulatory constraints it finds objectionable.

The declaration’s significance is political, not legal. It does not bind signatories to specific policies. What it does is establish India as a convening authority in global AI governance - a role that carries real weight in how standards, interoperability norms, and regulatory templates propagate through the 70-plus nations that signed.

The Architecture: Pax Silica and the Trade Dimension

The most consequential structural development at the summit received less coverage than the investment headlines but may prove more durable. Representatives of the U.S. and Indian governments signed the Pax Silica agreement - a U.S.-led initiative launched by the Trump administration aimed at securing the global supply chain for silicon-based technologies. The signing came alongside a framework for an interim US-India trade deal that would lower tariffs and increase economic cooperation between the two countries.

Read together, these agreements sketch the architecture of a new alignment. The U.S. is actively constructing a technology supply chain that excludes China and includes trusted partners - India chief among them. India’s combination of engineering talent, domestic market scale, and democratic governance makes it uniquely attractive as a Pax Silica anchor. The $18 billion in chip manufacturing projects that India has already approved are the industrial foundation that makes this alignment credible rather than aspirational.

Microsoft President Brad Smith put the strategic logic plainly in an interview at the summit: “If you look at the engineering talent, you quickly conclude India too can be a place where models are developed.” He predicted that “a variety of different DeepSeek moments” would come from India in the future - a reference to the Chinese AI lab’s surprise competitive breakthrough earlier this year. The comparison is deliberate. DeepSeek demonstrated that frontier AI development need not be concentrated in Silicon Valley. Smith is suggesting India could produce its own version of that disruption.

The Skeptics Are Not Wrong

The gap between summit promises and ground-level reality in India is wide enough to be honest about. Udith Sikand, a senior emerging markets analyst, offered a corrective at the margins of the event: “India is making splashy attempts to kickstart its belated AI push, but it is doing so primarily by offering headline-grabbing sops without addressing many of the underlying difficulties of actually doing business in India.” The observation is not uncharitable. India’s regulatory environment, logistics infrastructure, power grid reliability, and venture capital ecosystem all lag the ambitions being announced from the stage at Bharat Mandapam.

Anirudh Suri, founding partner of the India Internet Fund, identified the specific gap that matters most: “What we’ve not maybe seen as much of right now is venture capital and private equity money to come in to invest in Indian entrepreneurs in the AI space.” The infrastructure commitments from Reliance and Adani build data centers. The Microsoft and Nvidia partnerships wire those centers with compute. But the layer that converts compute into competitive AI companies - patient venture capital willing to fund Indian AI founders through long development cycles - remains underdeveloped.

The controversy that surrounded the summit offered a more pointed illustration of the gap between aspiration and execution. Bill Gates withdrew from the event after public backlash over his past relationship with Jeffrey Epstein - a reminder that India’s tech diplomacy operates in the same information environment as everyone else, subject to the same reputational dynamics. An Indian university was criticized after falsely claiming it had invented a commercially available Chinese-made robot dog. The story was minor in itself but symptomatic of a broader pressure to demonstrate AI capability that India does not yet possess at scale.

Why This Moment Is Different

Skepticism about India’s technology ambitions has a long and often well-founded history. The question is whether this summit represents another cycle of announcement followed by underdelivery, or whether something structurally different is happening.

The argument that something has changed rests on three factors. First, the capital commitments are anchored in genuine business logic rather than pure nationalism. Reliance and Adani are not making $110 billion and $100 billion bets as charity - they are responding to a domestic market of 1.4 billion people entering the digital economy at precisely the moment when AI is becoming the infrastructure layer of that economy. Second, the US-India alignment embodied in Pax Silica creates external structural pressure that previous Indian tech pushes lacked. When the United States is actively building India into its technology supply chain, the incentive for follow-through on both sides is real. Third, global AI infrastructure demand is growing faster than the established ecosystems can supply it. The hyperscalers need new data center capacity. India has land, renewable energy, and the political will to permit construction at the required scale.

None of this guarantees execution. Promises made at summits have a poor track record. The $250 billion will not materialize if India’s power grid cannot support it, if data center permitting stalls, or if the venture ecosystem does not develop the companies that actually use the compute being built.

But for the first time in the short history of global AI diplomacy, a country other than the United States or China has positioned itself not as a participant in someone else’s AI order, but as a shaping force in what that order will become. Whether India can sustain that positioning through the unglamorous work of building - the permits, the power lines, the policy reforms, the patient capital - is the question that will define the next decade. The summit answered nothing about that. It only clarified what is at stake if India gets it right.