New Delhi grew to become the primary metropolis within the Global South to host a worldwide synthetic intelligence summit when the India AI Impact Summit 2026 opened at Bharat Mandapam on February 16. The gathering—the fourth in an annual collection that started at Bletchley Park in 2023, moved to Seoul in 2024, after which to Paris in 2025—ran for six days, drew delegates from greater than 118 international locations, and registered over 5,00,000 attendees, making it the biggest such occasion up to now.
The summit got here with a string of main bulletins: a US-India strategic know-how alliance, a brand new home AI governance imaginative and prescient, billion-dollar funding pledges from Indian conglomerates, and the adoption of a multi-nation declaration backed by 88 international locations. It additionally got here with its share of logistical embarrassment—gridlocked streets, blocked delegates, a college expo stall evicted for passing off a Chinese-made robotic as an Indian innovation, and the last-minute withdrawal of Microsoft co-founder Bill Gates.
The summit was inaugurated by Prime Minister Narendra Modi on February 19. French President Emmanuel Macron and UN Secretary-General António Guterres each addressed the opening ceremony.
Modi used the event to announce India’s MANAV Vision, a framework for AI governance whose acronym—drawn from the Hindi phrase for human—stands for Moral and moral programs, Accountable governance, National sovereignty, Accessible and inclusive know-how, and Valid and bonafide programs.
“MANAV means human,” Modi mentioned. “M stands for moral and ethical systems: AI should be based on ethical guidance. A stands for accountable governance: transparent rules, robust oversight. N stands for national sovereignty: whose data, his right. A stands for accessible and inclusive: AI should be a multiplier, not a monopoly. V stands for valid and legitimate: AI should be lawful and verifiable.”
He known as for AI to be made extra broadly accessible: “We must democratise AI. It must become a tool for inclusion and empowerment, particularly for the Global South.” He additionally argued towards AI monopolies, saying India believed know-how would “only truly benefit the world when it is shared, when open-source code becomes available.”
Modi warned that deepfakes and fabricated content material had been “destabilising open society” and burdened youngster security as a precedence: “The AI space should also be child-safe and family guided.” He invited world know-how leaders to “design and develop in India, deliver to the world.”
The summit concluded on February 21 with the adoption of the New Delhi Declaration on AI Impact, endorsed by 88 international locations and worldwide organisations together with the United States, China, the United Kingdom, Russia, Brazil, and Australia. The non-binding declaration is structured round seven pillars—termed ‘Chakras’—protecting democratising AI assets, financial progress and social good, safe and trusted AI, AI for science, social empowerment, human capital, and resilient and energy-efficient AI programs. It was the broadest multilateral consensus on AI up to now.
Yet the occasion was not with out dissent. A TechPolicy.Press evaluation argued that the summit’s construction gave multinational firms parity with sovereign governments within the plenary classes, whereas civil society teams, labour leaders, and human rights defenders had no equal high-level platform. Amnesty International and others mentioned the summit prevented onerous conversations about dangerous AI use inside India itself. The US delegation, led by White House Office of Science and Technology Policy director Michael Kratsios, explicitly rejected any binding world governance framework. “We totally reject global governance of AI,” Kratsios mentioned on the summit’s closing day.
India joins Pax Silica
The summit’s most consequential diplomatic growth got here on February 20, when India formally signed the Pax Silica Declaration—a US-led strategic alliance geared toward constructing resilient provide chains for crucial minerals, semiconductors, and AI infrastructure. The signing occurred on the sidelines of the summit. Union Minister Ashwini Vaishnaw signed for India; US Under Secretary of State for Economic Growth, Energy, and the Environment Jacob Helberg, and US Ambassador to India Sergio Gor, signed for the United States.
India additionally signed a Joint Statement on the ‘India-US AI Opportunity Partnership’ as a bilateral addendum to the declaration.
Also Read | A borrowed robot exposes a republic of managed untruths
Pax Silica was launched on December 12, 2025, at a summit in Washington. Its founding declaration commits member nations to cooperation throughout the total know-how stack—from uncooked supplies by semiconductor fabrication to AI deployment infrastructure—with the said purpose of decreasing overconcentration in provide chains and stopping financial coercion. Current members embody Australia, Greece, Israel, Japan, Qatar, South Korea, Singapore, the Netherlands, the United Arab Emirates, and the United Kingdom.
India was not a part of the unique group. Its absence had fuelled hypothesis about commerce tensions with Washington; these tensions eased after India and the US reached an interim commerce deal final month that lowered US tariffs on Indian imports from 25 per cent to 18 per cent, and after India agreed to cease shopping for discounted Russian crude oil. Helberg had maintained on the time of the alliance’s launch that member choice was based mostly on supply-chain roles, not diplomatic friction.
Vaishnaw pointed to India’s semiconductor capability as the premise for its entry: “Our semiconductor industry has evolved well in the last decade. Today, we are designing 2-nanometer chips. It is very important to position India as a trusted country the entire world would like to partner with.”
Gor was extra direct in regards to the alliance’s underlying logic: “Pax Silica is about whether free societies will control the commanding heights of the global economy.”
India at the moment has 10 semiconductor crops at varied levels of growth, with industrial manufacturing anticipated quickly.
Investments, offers, and a few unease
The summit generated a rare quantity of funding bulletins, although analysts had been fast to notice that pledges usually are not the identical as disbursements, and that US tech giants are anticipated to spend over $630 billion on AI globally in 2026 alone—dwarfing India’s introduced totals.
The two largest home pledges got here from Reliance Industries and the Adani Group. Reliance chairman Mukesh Ambani dedicated Rs 10 lakh crore (roughly $110 billion) over seven years to construct AI and knowledge infrastructure, centred on a multi-gigawatt knowledge centre in Jamnagar, Gujarat, with over 120 MW anticipated to come back on-line within the second half of 2026. Adani dedicated $100 billion by 2035 to construct renewable-energy-powered hyperscale knowledge centres, projecting that this could draw an extra $150 billion in associated industries.
On the know-how firm aspect, OpenAI signed a partnership with the Tata Group to construct 100 MW of AI infrastructure in India, scalable to 1 GW, as a part of the worldwide Stargate initiative. Tata Consultancy Services additionally grew to become OpenAI’s first buyer for its knowledge centre unit. Anthropic opened a brand new workplace in Bengaluru—its second in Asia after Tokyo—and introduced a partnership with Infosys to deploy Claude fashions in Indian enterprises, starting with the telecommunications sector. Anthropic Chief Executive Officer (CEO) Dario Amodei described India as having “an absolutely central role” in shaping the way forward for AI.
OpenAI CEO Sam Altman meets Prime Minister Narendra Modi on the sidelines of the AI Impact Summit, in New Delhi, on February 20, 2026.
| Photo Credit:
ANI
Microsoft mentioned it was on target to speculate $50 billion in AI throughout the Global South by the tip of the last decade. Google CEO Sundar Pichai introduced a brand new subsea cable initiative to enhance AI connectivity between India and the US, and mentioned his firm was establishing an AI hub in Visakhapatnam.
BharatGen—India’s first government-funded giant language mannequin initiative, inbuilt collaboration with the India Today Group and the Ministry of Electronics and Information Technology—launched Param2, a 17-billion-parameter mannequin able to working throughout 22 Indian languages. At the summit, BharatGen additionally unveiled Sutra, an AI-powered information anchor designed to synthesise real-time coverage discussions into structured information stories in a number of languages. Sarvam AI launched a collection of recent fashions, together with a imaginative and prescient mannequin for optical character recognition, a dubbing mannequin, and a speech-to-text mannequin, and teased good glasses underneath the identify Sarvam Kaze.
Mumbai-based AI firm Fractal launched Vaidya 2.0, a healthcare reasoning mannequin that the corporate claims scores 50.1 on OpenAI’s HealthBench (onerous) benchmark—which, if verified by impartial reviewers, would make it the primary AI mannequin to exceed 50 on that measure, forward of OpenAI’s GPT-5 and Google’s Gemini Pro 3. The mannequin is designed to assist emergency triage, symptom checking, and end-to-end affected person journey administration. Fractal is a accomplice firm underneath India’s Rs 10,300-crore India AI Mission. The benchmark declare has not been independently verified on the time of writing.
Despite the size of those bulletins, critics warned of structural limits. India nonetheless lacks frontier computing infrastructure, chip manufacturing, and the large-scale knowledge centres wanted to coach aggressive fashions. Some researchers additionally cautioned that India risked turning into a ‘data colony’—supplying coaching knowledge and expert labour whereas international corporations retain possession of essentially the most priceless AI platforms. Startup funding in India fell in 2025, and a number of other founders flagged gradual authorities execution, together with delays in a deliberate $1.1-billion AI enterprise fund.
Water was one other concern raised on the summit, and outdoors it. A medium-sized knowledge centre can eat roughly 11 lakh litres of water per day. India faces acute and worsening water stress; Bengaluru, the place dozens of knowledge centres are already working, experiences extreme city water cuts every summer time. Modi’s invitation for the world’s knowledge to “reside in India” drew rapid pushback from environmental quarters.
The UN, governance, and a clumsy picture
Guterres delivered the summit’s sharpest line. Addressing the opening session, he warned that the way forward for AI “cannot be decided by a handful of countries—or left to the whims of a few billionaires.” He known as on know-how corporations to assist a $3-billion world fund to make sure open entry to AI for all international locations. He additionally introduced that the UN’s Independent International Scientific Panel on AI—comprising 40 consultants—had been formally appointed to assist evidence-based policymaking, and {that a} world dialogue on AI governance would maintain its first session in Geneva in July.
Macron emphasised Europe’s place as each a regulatory and an innovation area: “Europe is not blindly focused on regulation—Europe is a space for innovation and investment, but it is a safe space.”
Guterres’ warning about billionaires landed with explicit resonance on a day when Microsoft co-founder Bill Gates withdrew from a scheduled keynote simply hours earlier than he was because of converse. The Gates Foundation mentioned the choice was made “to ensure the focus remains on the AI Summit’s key priorities,” with out elaborating. Gates was travelling in India and had been confirmed as a speaker lower than 48 hours earlier. The withdrawal got here as renewed consideration targeted on his previous ties to the late intercourse offender Jeffrey Epstein, following the discharge of US Department of Justice information in January. Gates has not been accused of wrongdoing by any of Epstein’s victims. A Gates Foundation official spoke in his place. NVIDIA CEO Jensen Huang additionally cancelled his look, leaving two of the summit’s most anticipated audio system absent on its highest-profile day.
Also Read | AI in India is trapped in culture wars
The day additionally produced a second that unfold quickly on social media: when Modi prompted the know-how executives on stage to affix fingers in a gesture of solidarity, all complied—besides OpenAI CEO Sam Altman and Anthropic’s Amodei, who stood aspect by aspect however saved their fists raised individually. The two males lead the fiercest rivalry in AI; Altman had just lately known as Anthropic “dishonest” and “authoritarian” after the corporate ran Super Bowl commercials criticising OpenAI’s plan to introduce promoting into ChatGPT.
Also current on the summit however conspicuously subdued was China. The world’s second-largest AI energy despatched no senior authorities delegation; its absence coincided with Chinese New Year, although observers famous the timing additionally mirrored China’s broadly cautious method to multilateral AI boards dominated by the United States.
What the summit resolved—and what it left open
The New Delhi Declaration was the summit’s formal output. Unlike earlier summits, which targeted on AI security and danger, the declaration mirrored the host nation’s priorities: entry, growth, and the distribution of AI’s advantages to poorer international locations. Its seven pillars had been framed round alternative slightly than warning. It carried no binding obligations.
Jakob Mökander, director of science and know-how coverage on the Tony Blair Institute for Global Change, captured what many delegates felt: “Long term, it’s good for the world that AI is not just viewed as a race between the US and China, and I think that India is right now the player that most confidently says, ‘We reject this dynamic.’”
Last yr, India rose to 3rd place in Stanford University’s annual Global AI Vibrancy Tool rating of AI competitiveness—a leap that the summit’s organisers cited repeatedly. But because the The New York Times famous throughout the occasion, “India brims with tech talent but not the companies that command it.”
The expo was prolonged by a day and concluded on February 21, after drawing crowds far bigger than anticipated. The summit was the primary within the collection to be held outdoors a G7 nation. Whether its outputs translate into sturdy coverage or infrastructure—slightly than declarations and picture alternatives—is a query that can take significantly longer to reply.
With inputs from businesses.