India's AI Opportunity

The conversation about AI in India tends to go in one of two directions.

The first is catastrophist: AI will take millions of jobs, disrupt BPO, and leave India's workforce without the outsourcing engine that has driven its services growth.

The second is triumphalist: India's engineers will lead the AI revolution, Indian startups will build the next generation of AI companies, and the country will leapfrog from a developing economy to a tech superpower.

Both framings miss the more interesting and more important truth.

I am not making this argument from a distance. I grew up in Kalahandi, Odisha — one of the most underserved districts in India. I watched the internet arrive late, slowly, and unevenly. I watched Jio's entry in 2016 change something real: the young man who had been borrowing his cousin's phone to check train times suddenly had his own data plan. The information gap didn't close overnight, but the direction changed. What I saw firsthand is that connectivity, even imperfect connectivity, changes what people believe is possible — and that changes what they attempt.

AI is a more powerful version of that same mechanism. The question is whether it will reach the places that connectivity reached, or stop at the places where willingness to pay is highest.

The actual opportunity

India's AI opportunity is not primarily about building AI. It is about using AI to solve India-scale problems that no other country has had to solve.

India has 1.4 billion people. It has 22 official languages and hundreds of dialects. It has a civil service that processes billions of transactions annually, a healthcare system serving populations across radically different geographies, and an agricultural sector supporting hundreds of millions of small farmers.

Each of these is a problem of staggering complexity that has never been fully addressed — not because of a lack of intent, but because the tools didn't exist at the required scale.

AI is the first set of tools that has the potential to operate at this scale. Not AI as a product for the top percentile of the economy. AI as infrastructure for the whole economy.

The language opportunity

Consider one example: language.

India's linguistic diversity is one of its most beautiful and most challenging features. It has also been one of the most significant barriers to economic inclusion. Services designed for English or Hindi speakers have consistently failed to reach populations speaking Odia, Telugu, Marathi, or Bhojpuri.

Large language models, fine-tuned on Indian language corpora, have the potential to fundamentally change this. Not as a novelty — as infrastructure. The bureaucratic form in Odia. The agricultural advisory in Kannada. The legal consultation in Marathi. The medical referral pathway explained in Nagpuri to a patient who has never navigated a hospital before.

This is not a small opportunity. It is a transformational one. And it is one that no other country needs to solve in quite the same configuration — which means the companies that solve it will have built something with no direct precedent.

The healthcare gap

India has roughly one doctor for every 834 people. The WHO recommends one for every 250. That gap does not close by training more doctors — the pipeline is too slow. It closes, in part, by multiplying the effective reach of the doctors who exist.

AI-assisted diagnosis, clinical decision support, and patient triage at the community health worker level are not science fiction in India. They are already being piloted. The constraint is not imagination or even technology — it is distribution infrastructure to deploy these tools in the places that need them most, with the language support and connectivity assumptions that make them usable for a village ASHA worker rather than a hospital consultant.

The companies that crack this — that build AI health tools which function in rural Odisha as well as they function in urban Mumbai — will have built something worth far more than what is designed exclusively for India's premium tier.

The failure mode to avoid

The failure mode is building AI for the 100 million Indians who most resemble the customers of a Silicon Valley company, and calling it India's AI opportunity.

That is a market. A real market with real purchasing power and real problems worth solving. But it is not what makes India's AI opportunity distinctive.

What makes it distinctive is the scale of the underserved — the farmer who needs real-time weather and pest advisory in a language she speaks, the first-generation professional navigating formal institutions he was never trained to navigate, the small business owner whose financial life has moved to a phone but who has no access to credit or financial planning calibrated to his actual situation.

These are hard problems. The unit economics don't look attractive in the early phases. The deployment complexity is high.

But they are the problems that, once solved, create the most durable businesses — because the moats are operational and contextual, not just technical. Anyone can replicate a model. No one can quickly replicate the ground-level integration, the language training data, the trust earned in communities that have been failed by outside institutions before.

What this requires

Realizing this opportunity requires something India has historically underinvested in: the willingness to build infrastructure before the demand is fully visible.

The temptation is to build AI products for the premium Indian market — urban, English-speaking, already connected. That market is real and profitable. But it is not the transformational opportunity.

The transformational opportunity requires building for the next billion: accepting lower margins for longer, working with connectivity and device constraints that don't exist in the top-tier market, and maintaining a long time horizon that most venture capital structures don't naturally accommodate.

Some Indian companies will get this right. The ones that do will build some of the most defensible and consequential businesses of the next decade — and they will look, from the outside, like they took the harder path for no obvious reason. The reason will become obvious later.

Why India, specifically

I believe India will produce foundational AI companies — not despite its complexity, but because of it. The problems India has to solve are harder than the problems of smaller, more homogeneous markets. The solutions that work in India, at India's scale, are solutions that will work everywhere.

That is the opportunity. Not to copy what Silicon Valley is building. But to build for India and discover that you've built for the world.

I have seen a version of this in my own work. AI tools arrived in executive search in 2022 and 2023. The firms that used them to do the same work faster — to send more outreach, draft more templated messages, screen more CVs at speed — got marginal efficiency gains and a worse candidate experience. The firms that used them to do work that was previously impossible — to analyze why a pipeline was degrading, to identify patterns across hundreds of search outcomes, to reach candidates through channels that cold email never reached — found something genuinely new. The technology was the same. The question asked of it was different.

India's AI moment will follow the same logic. The companies that use AI to do the existing things faster will find the returns modest. The companies that use AI to do things that were previously structurally impossible — reach populations that no previous technology could serve, in languages and at price points that the existing market couldn't sustain — will build something that matters.