Last week, the AI Impact Summit in Delhi was mired in chaos because of mismanagement. In Beijing, meanwhile, a flawless humanoid robot show, inaugurated the Lunar New Year showcased the country’s cutting edge AI and robotics. The programme featured more than a dozen humanoid robots that gave a martial arts demonstration that included technically difficult innovations in robot coordination and fault recovery.In New Delhi, the news space was occupied by an exhibitor who showed a dog-robot, claiming it to be an indigenous product; it was actually a relabeled Chinese one.As for India, Prime Minister Modi himself noted that the summit showcased the extraordinary potential of AI in India. India has enormous talent manifested by the lakhs of Science, Technology, Engineering and Mathematics (STEM) graduates it produces and a deep diaspora embedded in leading AI labs. The country houses 20 % of the world’s chip design engineers and a workforce that supports their work. Now, the government is investing significantly in AI development.But though India is the home of leading IT companies like Infosys, Wipro and TCS, we have not been a leader in developing Large Language Models. Despite its STEM numbers, India’s AI industry growth is hindered by a significant shortage of specialised AI talent, limited computational infrastructure, as well as sustainability challenges arising from the massive energy and water required for data centres.Illustration: Pariplab Chakraborty.The AI summit was originally planned as a discussion on early stage safety issues, but it became a massive business oriented trade fair. The summit eventually served as a platform for India to pitch itself as a leader in AI deployment, infrastructure and investment.With global demand for AI compute capacity rising rapidly, countries are competing to attract large-scale data centre investments. Data centres and cloud infrastructure require high upfront capital investment, long gestation periods and sustained policy certainty. The 2026-27 Union government budget has initiated several steps to enhance India’s AI and data centre operations. Most important, there is a tax holiday proposed for eligible foreign cloud providers for using India based data centres for their global operations.Going by the announcements at the AI Impact Summit by industry majors, India has been successful in attracting massive investments, or at least the promise of them. But let us be clear, India is no AI superpower as yet. At the recent Davos meetings, IT minister Ashwini Vaishnaw took issue with the International Monetary Fund (IMF) chief who dismissed India as a second-level AI power. He cited a Stanford study that put India as being the number three power behind the US and China in “AI vibrancy” which recognises India as a rising AI hub due to its strong AI talent base, and growing AI investments.In reality, India’s AI industry is small with fewer AI researchers compared to even UK and France. As of now India relies heavily on adopting rather than creating foundational models. Moving from the high-level aspirational benchmarks to measurable, concrete impact on the ground remains the task ahead.An article in The Economist on the eve of the summit declared that as of now India was a “bystander in the race to develop frontier models.” To start with, India is strong on chip design, it is nowhere in the world of semiconductor manufacture. However, the journal notes that India’s data centre industry was now booming and its installed capacity had reached 1.3GW last year. That is small compared to the US (38.7GW) or China (9.5GW) but it is triple the figure of 2020.India’s stark position in the world of AI is evident from the fact that the number of high-end GPUs (graphics processing units) deployed in its national AI infrastructure is around 38,000. At the summit, Minister Vaishnaw said it would be boosted by another 20,000 in the coming weeks.But this compares poorly with the hundreds of thousands of GPU clusters in China and millions in the United States. Note that India is not in the list of 18 countries like Australia, Canada, Japan and UK, which have minimal restrictions in AI hardware imports from the US. India is with most of the other countries which are limited by US policy to importing just 50,000 GPUs till 2027. There is a third tier of countries like Russia, China, North Korea and Libya which face more formidable restrictions.Another major bottleneck to India’s AI ambitions is electricity. It is well known that AI data centres require massive uninterrupted reliable power which could be anywhere up to 100MW per facility. While India does have a high renewable energy potential , there are issues with inadequate infrastructure, grid stability and reliability. India has taken major strides in electricity generation in the past decade, with its installed capacity nearly doubling from 276GW in 2014-15 to nearly 500GW in 2025. Renewable capacity grew threefold to 227GW. Besides electricity is the requirement of immense amounts of water for cooling data centres which is an issue in an already water-stressed region like Bengaluru.India’s electricity generation pales before China’s which added 543GW in 2025 alone, and has an installed capacity of 3,891GW. The Chinese goal is to ensure a stable and abundant energy supply and limit fuel imports and thereby provide competitive advantage to Chinese industry.India does have plentiful data, but there are issues about its quality and accessibility. There are issues, too, with the poor quality of government data and the paucity of metadata necessary to understand data quality problems. India’s data tends to be siloed, unstructured, unlabeled and inaccessible, leading to inaccurate AI outputs. But the government push to localise data such as an RBI law to mandate that financial institutions retain client data within the country has encouraged data centre growth.A major problem India confronts is its ability to hold on to AI talent. Brain drain is a real problem and the best Indian AI talent is pulled away to Silicon Valley or the UK because Indian companies often are unable to offer the salaries demanded. In the case of start ups, there is also the tendency to leave the country to seek funding abroad. So India’s position is somewhat paradoxical. It has significant potential and has generated a great deal of momentum for its aspirations. But it has meaningful structural gaps.The AI Impact Summit has certainly played a role in focusing attention on the issue of AI in the country. But this has emphasised the business end of the subject. With tax exemptions and a huge pool of users, India offers a vast market for AI products. But this is more about profits, rather than the issues spoken of by Prime Minister Modi, encapsulated in the acronym MANAV – moral and ethical systems, accountable governance, national sovereignty, accessible and inclusive and valid and legitimate systems.The writer is a distinguished fellow with the Observer Research Foundation in Delhi.This piece was first published on The India Cable – a premium newsletter from The Wire – and has been updated and republished here. To subscribe to The India Cable, click here.