As the United States and China dominate the global artificial intelligence race, India still lacks publicly listed pure-play AI companies. However, the country’s startup ecosystem has reached a turning point that could change that, Rajan Anandan, Managing Director at Peak XV Partners, said at an Express Adda discussion on Monday.

“India is early in the AI game. 2026 is the year India entered the AI game,” Anandan said, adding that while global capital continues to favour markets with dedicated AI companies, India’s private innovation ecosystem tells a very different story.

The discussion featured Rajan Anandan and Sarvam AI Co-founder and CEO Pratyush Kumar in conversation with Anant Goenka, Executive Director of The Indian Express Group, and Soumyarendra Barik, Assistant Editor at The Indian Express, on India’s AI ambitions and its place in the global technology race.

Anandan said the absence of listed AI companies often creates the impression that India is falling behind. He pointed out that companies such as E2E Networks are rapidly transitioning towards AI, while larger firms like TCS are increasingly integrating AI into their businesses. He said, India’s early-stage startup ecosystem is witnessing significant AI innovation.

He said these startups are attracting talent, building AI products and laying the foundation for globally competitive companies.

Pratyush Kumar agreed that India has a unique opportunity despite the current dominance of the US and China. “There is no way India will not win in AI,” Kumar said. “Technology and innovation always gravitate to where there is demand,” he said. 

He said countries around the world are competing to build sovereign AI capabilities, but India’s growing demand for artificial intelligence gives it a significant long-term advantage.

Why can India compete in AI?

Responding to a question from Anant Goenka on what gives India the right to compete across the AI stack despite the technological leadership of the US and China, Kumar said artificial intelligence is unlike any previous technology because it combines frontier innovation with enormous economic value.

“If you plot technologies on two axes – economic value and frontier technology – AI is at the top of both,” he said.

He described AI as a “general purpose technology” that will reshape almost every sector, including government services, enterprise software, entertainment and workplaces.

According to Kumar, India’s biggest strengths are its talent, ambition and demand for AI solutions. “The future of work is going to look very different,” he said. “India will need AI in every walk of life,” he said. 

He said that building frontier AI models requires years of investment in engineering talent, research, computing infrastructure and capital.

Drawing a comparison with OpenAI, Kumar said the company started building large AI models in 2015 and spent years developing the required expertise. “We are only in our third year,” he said, referring to Sarvam AI’s efforts to build frontier AI models.

He said Sarvam began without prior experience in training frontier-scale models but gradually built capabilities by investing in data, GPU infrastructure and engineering talent. “There is no way India won’t be ultra-competitive in AI because the country’s demand will force us to be ultra-competitive,” Kumar said.

AI stack consists of five layers – power, chips, infrastructure, models and applications – and each layer requires sustained investment and technical expertise.

Does India need the world’s biggest AI models?

Both speakers argued that India does not need to copy the strategy followed by global frontier AI laboratories.

Anandan said the overwhelming majority of businesses do not require massive frontier AI models. “Ninety to 95% of what enterprises need can be done with smaller models,” he said.

He said smaller AI models require significantly less capital and are far more practical for enterprise adoption in India.

Kumar said Sarvam AI has adopted the same approach by building affordable AI models designed specifically for Indian users and businesses. “Our AI models are five times cheaper than GPT and nine times cheaper than Gemini Flash,” he said.

He said frontier AI laboratories are building one-size-fits-all models for the world, while Sarvam is developing cost-efficient AI models tailored to India’s languages and enterprise requirements.

Kumar also praised China’s rapid AI progress, saying it demonstrates how sustained investment in research, infrastructure and talent can transform a country’s technological capabilities. “China has a great model, and we should celebrate it more,” he said.

According to Kumar, once countries build sufficient talent, compute infrastructure and capital, they reach a stage where innovation becomes far more competitive. He said India has already started building that foundation.

Anandan said affordable electricity could become one of India’s biggest competitive advantages in AI. 

Recalling a recent visit to the United Kingdom, he said high electricity costs have become a major hurdle for AI development there. “If you’re in a nation with expensive power, you won’t have AI,” he said.

He added that India’s rapid expansion of solar power, renewable energy and energy storage over the next five years could significantly strengthen the country’s AI ecosystem as demand for computing infrastructure continues to rise.