India's Grand AI Gambit and How World's Biggest AI Summit Lurched Between Scandal, Spectacle, and Genuine Promise

When Prime Minister Narendra Modi stood before a packed Bharat Mandapam on 19 February 2026 to inaugurate the India AI Impact Summit, the symbolism was unmistakable. For the first time in the four-year history of international AI summits, a series that had previously cycled through Bletchley Park, Seoul, and Paris, the global conversation had arrived in the Global South. With French President Emmanuel Macron and United Nations Secretary-General António Guterres both present at the opening ceremony, New Delhi had, at least on paper, secured its place at the top table of the world's most consequential technology debate.

The summit, organised under the IndiaAI Mission by the Ministry of Electronics and Information Technology (MeiTY), was structured around three foundational pillars termed "Sutras", People, Planet, and Progress, and seven working groups spanning everything from democratising AI resources to climate resilience. Over 100 countries sent participants; 300-plus exhibitors filled ten thematic pavilions; and by the close of the first two days, India had set a Guinness World Record for the most pledges received for an AI responsibility campaign in 24 hours, with 250,946 valid pledges collected in partnership with Intel India, far exceeding the original target of 5,000.

The geopolitical optics were considerable. Google CEO Sundar Pichai announced a landmark partnership with Reliance Jio for new cloud clusters and a 50-megawatt renewable energy project in Rajasthan. Anthropic opened its Bengaluru office. Larsen & Toubro unveiled a proposed gigawatt-scale NVIDIA AI factory under the India AI Mission. Union Minister Ashwini Vaishnaw announced the addition of 20,000-plus GPUs to India's existing 38,000-unit cluster.

Yet the summit's most prominent absence spoke loudly. Microsoft co-founder Bill Gates, who had travelled to India and worked in Andhra Pradesh, withdrew just hours before he was scheduled to speak. The Gates Foundation cited a desire to "ensure the focus remains on the AI Summit's key priorities," coming as Gates faced renewed scrutiny over his association with the late sex offender Jeffrey Epstein. Another Foundation official stepped in. Separately, NVIDIA CEO Jensen Huang also cancelled his appearance, robbing the expo of two of its most-anticipated draws at a moment when India was keenest to demonstrate it could command the world's attention.

Modi's own address struck a note of grand industrial ambition: "Data centres will be massive job creators for youth; we invite the whole world's data to reside in India." It was a line that would soon attract significant pushback.

The Human Cost: Evictions, Chaos, and the People Left Outside the Gates

Long before the first keynote, the preparations for the summit extracted a price from those at the very bottom of Delhi's social order. In the weeks leading up to the event, major cleanup drives swept through the city. Homeless individuals who had sheltered beneath bridges and in makeshift cold-weather camps, many having sought refuge during the winter chill, were evicted overnight, beaten, and relocated to distant homeless shelters. At Nehru Place, where shelters had already exceeded capacity, and people were sleeping outside, the consequences were starkly personal.

The Municipal Corporation of Delhi issued an order stating that, "Keeping in view the winter season during the summit period, as well as the imperative to maintain public health, safety, cleanliness, and a dignified urban environment, it is essential that such homeless persons are shifted well in advance to appropriate shelter facilities, with due care for their welfare and basic needs." The language of dignity, it would appear, did not extend to those whose dignity was being most directly violated in its name.

Inside the venue, a different kind of chaos unfolded. The summit was overregistered beyond its own official capacity, resulting in scenes reminiscent of a poorly managed mela rather than an international technology conclave. On the first day, UPI payments were inexplicably barred at food stalls, forcing attendees to pay in cash or queue to load prepaid cards, a decision that provoked widespread anger and was hastily reversed on day two. Food was overpriced, water scarce, and toilet facilities overwhelmed, with several blocked entirely.

Delegate accounts were damning. Cyber lawyer Vratesh Parmar,who attended the summit and spoke to us at the venue, noted, "There were separate queues from separate entry points for people with and without bags. Someone with a laptop bag had to walk all the way around to a separate entry, which in some cases was right behind the building. Toilets were full, and a couple of them choked and blocked. I saw the official India AI mission page’s post. We understand difficulties in such events, but just taking defence will not work."

Parvathi Bakshi, a registered delegate, was equally pointed, "Maps and signage were of poor quality, 'West Wing' versus 'West Pavilion,' sometimes referred to as 'Meeting Room' versus 'Hall,' and the worst was the lack of a stall numbering system in the expo halls. Security was a joke. I didn't have a laptop but security staff were demanding to confiscate my notebook, saying 'Madam, from tomorrow, laptops and notebooks, nothing is allowed.' That's on the organisers to improve SOPs and coordination with police and security. The stampede risk at the West Wing hallway trying to access Rooms 4A, 4B, and 6 was a nightmare."

View this post on Instagram

Srikant Mohapatra,a tech startup developer who travelled from Bhubaneswar, recounted standing outside the venue for two to three hours on day one with no official communication, only to be told by Delhi Police to vacate the area until 6:30 pm. "On the third day, there was a rain problem; workers were working hard to dry the place, but open areas were still slippery, and one gentleman fell right in front of me," he said. "For an international summit with over 2.5 lakh people in attendance, management mistakes are not a joke."

Bloomberg separately reported that delegates were left stranded without food or water during a security lockdown ahead of the Prime Minister's visit on 19 February, a day on which entire sections of the expo were evacuated, and exhibitors lost critical showcase time with potential clients.

A Bengaluru-based startup, NeoSapien, suffered even more directly: founder Dhananjay Yadav returned after a PM-visit security clearance to find his company's AI-powered wearable devices missing from the booth. He posted on X, "Think about this — we paid for flights, accommodation, logistics and even the booth. Only to see our wearables disappear inside a high-security zone." Delhi Police later identified and arrested two contractual workers in connection with the theft, recovering the devices.

The Innovation Optics Problem: Fake Dogs, Real Shame, and the Startup Gold Nobody Saw

If mismanagement was the summit's administrative failure, the innovation scandals were its reputational one, and both Galgotias University and Wipro found themselves at the centre of a controversy that exposed uncomfortable truths about how India presents its technological ambitions.

On 18 February, a video broadcast by Doordarshan showed a Galgotias University professor introducing a four-legged robotic device at the institution's exhibition pavilion as "Orion", an in-house creation from the university's Centre of Excellence, built as part of a claimed Rs. 350-crore investment in its AI ecosystem. Within hours, social media users had identified the robot as the Unitree Go2, a commercially available quadruped manufactured by the Chinese company Unitree Robotics. The university was directed to vacate its stall by MeiTY Secretary S. Krishnan, who made clear the government did not want exhibitors showcasing items they had not developed. The university issued an apology, asserting the representative had been "ill-informed" and was not authorised to speak to the press, but the damage was done, and a communications professor was made the public face of the fallout.

pic.twitter.com/6locYz5QaL

— Galgotias University (@GalgotiasGU) February 18, 2026

Wipro also drew criticism for displaying what appeared to be the same Unitree Go2 model, branded internally as "TJ Robot", at its own pavilion. The company later clarified that it had never claimed to have built the hardware, and that the robot was being used to demonstrate AI software applications in enterprise and disaster-response contexts. The Unitree markings had, however, been removed from the device, and the distinction between software demonstration and hardware misrepresentation proved too subtle for a public already primed for scepticism. 

Tata's "AI weaving sensor" was similarly challenged. Attendees accused the exhibit of being basic automation and sensor coding bundled together with no genuine AI model involvement. "Not all automation is AI," one student from IIT Delhi, Sakshi Mehrortra, said pointedly, standing outside the stall. "The world has to learn that. I'm amazed that bigger companies are doing this."

The tragedy, as multiple observers noted, was that genuine innovation was present in abundance, it simply could not cut through the noise. xTerra Robotics showcased advanced robotic systems built in India. Svan's M2 robo-dog was constructed entirely from scratch domestically, in stark contrast to the imported models that stole headlines. Sarvam AI's Kaze smartglass, a homegrown competitor to foreign wearable AI brands, aiming for market readiness by May 2026, deserved significant attention. SarvamAI, one of the few Indian names operating at scale in the generative AI space, continued to represent what India could genuinely build.

Turns out, Sarvam Kaze comes with a smart ring for controlling the functionality of the smartglasses!

The built-in live Indian language translation feature is going to be a game-changer.

Big thanks to Sarvam for letting me do this hands-on impromptu demo. pic.twitter.com/EcCEz6JPDM

— Caleb (@caleb_friesen) February 18, 2026

Amrit Pritam, a Galgotias student who attended the summit under a separate organisation, offered perhaps the most honest self-reckoning, "What happened to Neha ma'am is not fair. The problem is much larger than what they are pretending it to be. It is not mere miscommunication; it is optics. We do have centres of excellence, we do have large tech investments and very bright students, many of whom will make incredible startups. But out of the 1089-odd patents we filed, none got approved due to lack of ingenuity. In a world of AI where information access is so easy, it is a tough battle. That should be the conversation, and not how optics are placed as the biggest, brightest, best. We are not, and we should rather accept that."

MeiTY's Secretary acknowledged in a post-incident statement that stricter norms were needed for exhibits at future expos, a quiet concession that the gatekeeping had failed.

Climate, Data, and the Harder Conversations the Summit Tried to Have

Beneath the controversies, some of the most substantive discussions at the summit centred on whether India's AI ambitions were ecologically coherent, and the answers were not particularly reassuring.

Modi's invitation for the world's data to "reside in India," framed as a job-creation opportunity, drew immediate concern from environmental quarters. A medium-sized data centre can consume approximately 11 lakh litres of water per day. India, a country facing acute and worsening water stress, already struggles every summer with urban water cuts, and nowhere is the tension more visible than in Bengaluru, where dozens of data centres have established themselves in what is often called India's Silicon Valley.

Dr Rajendra Singh, an environmentalist specialising in groundwater conservation, was direct, "Every summer there is news of water shortage. Water cuts are most frequent in urban areas. Bengaluru, around which multiple data centres have popped up thanks to the Silicon Valley of India label, faces water cuts so critical that they become a major civic issue. Sixty years we have not yet solved the Kaveri issue, and here we are speaking about massive data centres and how we want the world to do more exploitation."

It was precisely this tension between digital ambition and environmental reality that Dr Arunabha Ghosh, CEO of the Council on Energy, Environment and Water (CEEW) and Chair of the Expert Engagement Group on Climate and AI for the summit, sought to address. In his remarks, he said, "The AI Impact Summit hosted by India, the first one in the Global South, signals a shift: countries most exposed to climate risk are beginning to shape the rules of the next technological transition. The real value of artificial intelligence lies in managing complexity. Climate change is a complex and wicked problem. To spur climate action, AI can be leveraged to balance renewable-heavy power systems, improve flood forecasting, and strengthen agricultural resilience at scale.

"For India, with expanding digital public infrastructure at the population scale, this creates an opportunity to embed climate intelligence directly into development decisions. But progress must come with safeguards. Data centres consume energy and water, yet governance frameworks lag. Responsible AI for climate, therefore, means transparent environmental reporting, right-sizing models, careful infrastructure siting, and keeping humans firmly in the loop. How we apply AI for climate must reflect our choices, and we must build not just AI agents but also invest in the human agency to do so."

It was a formulation that offered genuine intellectual coherence to a summit that had, in many moments, struggled to present a credible face to the world. Google's announcement of a Centre for Climate Technology in India, in collaboration with the Principal Scientific Adviser to the Government of India, pointed in the right direction. So did the broader international conversation about AI governance that the summit formally advanced, the United States, notably, arrived with an agenda centred on "domination" rather than cooperation, framing AI as a geopolitical contest against China, a posture that sat uncomfortably alongside the summit's stated goals of inclusive, people-centric development.

The UN Secretary-General António Guterres captured the wider stakes plainly: the future of AI, he warned, could not be left to the "whims of a few billionaires." It was a statement that landed with particular resonance at a summit where multinational corporations had been accorded parity with sovereign governments in the plenary structure, while civil society and labour leaders had no equivalent high-level platform.

India's AI Impact Summit 2026 will be remembered as a moment of genuine historical significance, the first time the Global South claimed the chair at AI's highest table. Whether the chair was ready for the occupant, and whether the occupant was ready for the responsibilities that came with it, are the questions that will take longer to answer. What is clear is that between the evicted rose-seller at Nehru Place and the Google CEO announcing billion-dollar partnerships inside Bharat Mandapam, the distance was not merely one of wealth. It was a distance between whose futures were being shaped, and whose were merely being tidied out of the way.