AI Summit: Here's why Lightspeed’s Bejul Somaia sees India’s AI future in applications, not infrastructure

Somaia also argued that AI could ease one of India’s most enduring constraints: access to top-tier talent.

• India's AI opportunity lies in application layer, not core models
• Falling AI costs enable wider access to healthcare, education
• AI lets smaller teams build startups faster, changing competition

Did our AI summary help?

India’s biggest artificial intelligence opportunity will emerge in the application layer rather than in building large foundation models, according to Bejul Somaia, Managing Director at Lightspeed Venture Partners.

Speaking on the fourth day of the India AI Impact Summit on February 19, Somaia said that while much of the global focus remains on developing large language models and core AI infrastructure, India’s strengths lie elsewhere.

“While the foundation models, the large language models, the reasoning engines are largely being built outside India, with the exception of Sarvam, this will not be the primary area of opportunity in India. The primary opportunity area here is in the application layer,” he said.

India’s AI moment mirrors early internet era

Drawing parallels with the rise of the Indian consumer internet in the late 2000s, Somaia said the country is at a similar inflection point today — but with a much narrower window to act.

“We were not just Indian versions of American ideas. They were fundamentally different companies built for a fundamentally different context,” he said, referring to the first generation of India-focused internet startups.

That experience, he argued, offers a blueprint for how India should approach AI — by reimagining use cases for local workflows, languages and regulatory environments rather than replicating global products.

“The history of the Indian consumer internet is a history of taking global ideas and rebuilding them from first principles for a market that global players fundamentally misunderstand,” he said.

Falling AI costs changing the equation

A key shift enabling this opportunity is the rapid decline in the cost of AI deployment, which Somaia described as unprecedented.

“The cost of intelligence has collapsed. It hasn’t declined — it has collapsed,” he said, noting that AI inference that once cost hundreds of dollars now costs fractions of a cent.

Unlike earlier technology waves that needed years to solve affordability challenges in India, AI economics are already viable, he said, removing a major barrier to mass adoption.

“Every technology wave that came before — smartphones, broadband, software — had to solve the affordability problem to unlock the India market… AI is different,” Somaia said.

Levelling access to expertise

Somaia said the drop in AI costs could significantly expand access to healthcare and education — two sectors where India faces persistent access gaps.

AI, he said, can make “the intelligence of the best diagnosticians in the world accessible to a primary care provider in a third-tier city” and enable large-scale personalised learning systems that were previously impossible to build.

“The opportunity is not just to digitize an existing system. It is to reimagine what healthcare delivery looks like when intelligence is abundant and geography is irrelevant,” he added.

AI reshaping how startups are built

Somaia also argued that AI could ease one of India’s most enduring constraints: access to top-tier talent.

“When intelligence becomes abundant, when a founding team of five can do the work that previously required fifty, the talent bottleneck changes in character,” he said.

This shift, he noted, allows smaller teams to build with far greater leverage, fundamentally altering how startups scale and compete.

A call for speed, not parity

Rather than measuring India’s progress against current global leaders, Somaia urged founders and policymakers to focus on momentum.

“What matters is not where you are, but how fast you are moving. Scale is a snapshot, but the slope, or trajectory, is the story,” he said.

He framed the current moment as analogous to 2008, when early Indian internet entrepreneurs built for a market that had not yet fully arrived — but warned that AI adoption cycles are moving far faster.

“The question is simply who will be fast enough to build the right applications before this window closes,” Somaia said.

As AI capabilities become cheaper and more widely available, he concluded, India’s competitive advantage will depend less on replicating foundational technology and more on translating it into real-world use cases.

“Everyone listening to this is not just a spectator,” he said. “You are the protagonists.”