India’s AI summit: Where big ambitions met hard realities

NEW DELHI – India’s

AI Impact Summit 2026

was billed as a coming-of-age moment for the country, as it seeks to enter the global artificial intelligence race dominated by the rivalry between the US and China.

Instead, the summit presented a far more complex picture, underlining both the country’s ambition to innovate and provide leadership, and its constraints in terms of infrastructure bottlenecks and dependence on foreign technology.

Logistical problems took some shine off the event. Glitches in accreditation, last-minute schedule changes, VIP movement cutting off access for participants and exhibitors, long queues within the venue and traffic jams dampened the summit’s sheen.

An Indian university was evicted from the summit in a

viral incident

after one of its officials falsely claimed to the Indian media that a Chinese-made robotic dog was its own invention.

Despite all this, the message from the summit was that India is open for AI business and is more than a market for tech giants.

“The whole intent and purpose was to really put on display India’s seriousness in the field of AI – both with regard to the creation of technology and the development of products, and the application of AI for tackling persistent larger economic problems and human development problems,” said Mr Rentala Chandrashekhar, chairman of the Centre for The Digital Future.

Despite “a lot of noise”, the message that India is pursuing these objectives came through, said Mr Chandrashekhar, who was formerly the top official at the Ministry of Electronics and Information Technology.

Announcements of deals worth billions of dollars

reinforced this message.

Reliance Industries chairman Mukesh Ambani, Asia’s richest man, pledged investments of around US$110 billion (S$139 billion) over the next seven years to build AI and data infrastructure in India.

Indian conglomerate Adani Enterprises said it would invest US$100 billion to build renewable-powered, AI-ready data centres by 2035.

Partnerships were established, with Tata Consultancy Services partnering OpenAI to build AI infrastructure and Infosys entering into a partnership with US-based Anthropic to provide AI solutions to companies across telecommunications, financial services, manufacturing and software development.

Meanwhile, Nvidia unveiled tie-ups with three Indian cloud computing providers to provide advanced processors for data centres that can train and run AI systems.

All this indicates that India has entered the AI race, said former minister Rajeev Chandrasekhar, who asserted: “The capacity and capability that exist in the Indian research and innovation ecosystems are solid.”

Mr Chandrasekhar, a leader of the ruling Bharatiya Janata Party and former minister of state for electronics and information technology, added: “We may have been a little late to the AI race, and we will need to do a lot of catching up.

“It is not easy because of the sort of walls and the moats that have been built by the US and the Chinese around their own products and their LLMs. But that is a race that we are going to have to run.”

LLMs, or large language models, such as OpenAI’s ChatGPT, Google’s Gemini and China’s DeepSeek, are AI systems designed to understand, process and generate human language.

India is ChatGPT’s second-biggest market, with 100 million weekly users.

But competition is heating up, with Indian start-up Sarvam AI hoping to give generative AI tools such as ChatGPT and Claude a run for their money.

The company unveiled two models built and trained with local datasets, making them more culturally attuned to Indians in terms of languages and culture.

Poised for the AI leap

“The power in India is the population and the amount of data that you can drive to drive (AI) models. Think about it. AI has nothing without good data training it,” said Ms Vanessa Smith, a speaker at the summit and chief corporate affairs officer at American company ServiceNow, which provides a cloud-based platform for automated business workflows.

India also has a proven track record of building IT services, which can be leveraged into AI systems. Any application developed in India can be exported to countries in the Global South, given the country’s diverse linguistic and cultural diversity, experts said.

At the summit, Indian Prime Minister Narendra Modi spoke of how technology developed in India would help other countries as well, maintaining: “Any AI model that succeeds in India can be deployed anywhere in the world.”

Held in New Delhi from Feb 16 to 21, the summit is the fourth of its kind after similar annual conferences in Britain, South Korea and France to discuss the problems and opportunities posed by AI.

Among its 70,000-plus attendees were government delegations,

world leaders

such as French President Emmanuel Macron, and industry players, including top technology executives like OpenAI’s Mr Sam Altman and Google’s Mr Sundar Pichai.

India is hosting the gathering at a time when AI adoption has accelerated worldwide, raising questions over safety, job losses and ethics amid the rise of agentic AI systems. Advances in AI have already rendered a number of jobs, including traditional white-collar roles, redundant.

Hurdles to cross

A key aim of the summit was to “globally bring in the perspective of the Global South into a space that is increasingly dominated by big powers (US and China)”, said Professor Harsh V. Pant, vice-president at the Observer Research Foundation think-tank in New Delhi.

He said a key message from the summit was that AI needs to be more inclusive and democratic, and “that the technological divide that we have seen with regard to other technology should not be replicated when it comes to AI”.

A significant challenge for India is that the AI ecosystem remains dominated by the US and China, which have sought to protect their AI models.

India has been drawing closer to the US in recent years, even as it remains distrustful of China, with which it has a disputed border.

The South Asian nation’s nascent semiconductor chip industry leaves it still reliant on foreign AI stacks – the software, tools and models an AI system is built on – and processors.

India on Feb 20 joined the US-led Pax Silica initiative aimed at strengthening resilient supply chains for critical minerals and AI – a move seen as India trying to get around China’s dominance in these areas.

But even as it deepens technological cooperation, India faces the challenge of maintaining strategic autonomy – and that includes navigating pressures from its closest partners.

At a Tony Blair Institute event during the summit, Mr Sriram Krishnan, senior White House policy adviser on AI, was explicit about how the US wanted India to continue using American technology.

Mr Krishnan noted that while Indian companies should localise applications, “at the end of the day, we want the American AI stack to be the bedrock that everyone builds on”.

India also needs to move fast to boost its infrastructure to match its AI ambitions. For one thing, vast amounts of energy and water resources will be needed to power AI data centres.

“India’s doing pretty well on its energy needs right now. But the amount of energy that data centres and AI centres typically tend to use is pretty high. So working towards making sure (energy needs are met) is going to be important,” said Mr Vivek Agarwal, country director for India at the Tony Blair Institute, noting that India’s recent efforts to push nuclear energy are a “big step” in that direction.

The summit did succeed in giving India a louder voice in the global conversation on AI technology, while highlighting its potential in the sector.

“We had the right people in the room (at the summit); you have the private sector, the academy and the government, and not just, you know, advanced countries, but also developing countries,” Mr Agarwal noted.

“The US is the front runner with Nvidia and others (in the AI race); China will probably catch up soon. And I think the third country to do that will likely be India. And I think nobody in the government or otherwise will disagree that it will take a lot of effort.”