India AI Impact Summit: Dr Vishal Sikka argues that India can mirror the Green Revolution to democratise AI for a billion people

The India AI Impact Summit 2026, being held from February 16 to 20 at Bharat Mandapam in New Delhi, features participation from more than 110 countries and 30 international organisations, including about 20 heads of state or government and almost 45 ministers. On day four of the Summit, key global AI figures such as Sundar Pichai, CEO, Google, Sam Altman, CEO, OpenAI, Dario Amodei, CEO of Anthropic, and prominent heads of state, including French President Emmanuel Macron, will address the audience.

Speaking at a pivotal keynote on the third day of the Summit, on February 18, Dr Vishal Sikka, Founder and CEO of Vianai Systems, addressed a central theme of democratisation of artificial intelligence (AI) resources, diving into an urgent concern in India’s AI trajectory. Sikka offered fascinating insights into the underlying challenge of ensuring equitable access to this transformative technology amid global disparities in compute power and data ownership. This topic matters profoundly because India, with its vast population and burgeoning tech talent, must ensure resources such as GPUs and foundational models become widely available to startups, schools, and individuals.

AI as collective knowledge
“Large language models [LLMs] are, in a sense, a compression of a large amount of knowledge,” he said. “The compression of a large amount of knowledge. A large amount of human knowledge gets stored in some way.” He explained that these models, though enormous, remain smaller than the vast data they process, offering approximate outputs that sometimes veer into hallucination. This ability to “regurgitate that in a very conversational way” enables authentic interactions in languages such as Gujarati, Bengali, or English, raising critical questions about ownership. “Whose data is that?” Sikka asked, noting concerns over copyrights while asserting that AI represents “a collective compression of the world’s knowledge” belonging to all, even if controlled by a few frontier labs.


Lessons from the Green Revolution
Sikka drew a compelling parallel to India’s Green Revolution to advocate for mass AI literacy. He recounted a childhood memory from around age two or three, during the 1971 war that birthed Bangladesh, when his parents fretted over a wheat-laden ship delayed by the ceasefire. “My parents were genuinely worried whether we would have food to eat or not,” he shared. Yet within a generation, India transformed from food scarcity to exporting wheat and butter through initiatives by companies like Amul.

“If we can do that, you better believe we can bring a billion people into the world of AI,” Sikka declared. He praised the government’s vision, calling India the “leading country from the point of view of the government’s involvement in AI,” citing the India AI Impact Summit itself as testimony to this scale.

AI’s energy hurdles and the human brain
Sikka made a sharp distinction between the current reality of AI’s massive energy and compute needs and the inspiring efficiency of the human brain, to argue for a two-pronged path forward.

Right now, training even one major model, such as the original ChatGPT (GPT-3 based), “took something like 1.3 gigawatt hours of training,” as he noted, equivalent to powering thousands of homes for a year.

The Government of India currently has 38,000 graphics processing units (GPUs), with another 20,000 more imminent to take the count to 58,000 within weeks at Rs 65 an hour, which is ambitious but, Sikka pointed out, pales against giants such as OpenAI’s million GPUs last year alone. Sikka stressed securing this scarce compute for startups and schools as a short-term must, adding that the government is doing this “better than anybody else.”

But his deeper point, that we must invent next-generation AI that’s as power-thrifty as our human brains, is forward-looking. “When we operate on 2,000 calories in a day, that's like 100 watts, like one of these light bulbs […] and out of that, our brain is taking up 20 watts,” Sikka said. That 20 watts, which is less than that of a phone in sleep mode, powers complex multisensory processing from seeing and speaking to gesturing and reasoning, far beyond today’s models.

“Why don’t we invent the next generation of AI, which is much more power efficient? When we know it is possible, just look at yourself,” he urged. This dual strategy, making use of available compute while pioneering efficient systems, lies at the heart of true AI democratisation.

Opportunities over job fears
On talent and displacement fears amid brain drain and the global capability centre (GCC) wave, especially with talent flows shifting back home under President Trump, Sikka rejected panic.

He recalled early television’s awkward moments with newsreaders holding papers, akin to “radio with visuals”, before live cricket and MTV showed how dynamic TV could be. He explained that AI, today, mirrors this, with fixation on automating old tasks overlooking tomorrow’s breakthroughs.

AI unlocks “wondrous applications,” Sikka argued, citing the example of a distribution client using AI-led scenario analysis to see “What if prices rise here and demand falls there?”, resulting in shuttering a factory in days, not the year it would have taken consultants via backward-looking quarterly reviews. “Don’t focus on what is it that it takes away, focus on what is possible,” Sikka advised, urging a shift from disruption to forward innovation, with AI agents on the horizon to amplify India’s talent.

Sikka’s keynote ended on a high note of optimism, inspiring the audience with a vision of AI as an empowering force for India’s billion-plus population. By emulating past triumphs such as the Green Revolution and articulating democratisation as a top priority, he argued, India can lead the world in harnessing AI for prosperity.