AI key to India's $30 trillion economic goal: Bharat1.AI's Umakant Soni

A three-to-five-year window will decide whether India builds its own AI ecosystem or becomes a consumer, says Bharat1.AI chief Umakant Soni, warning of growth risks

India must build domestic artificial intelligence capability while leveraging its demographic dividend to avoid restricting the economy’s size to $8 trillion by 2047, one-third of the government stated goal of $30 trillion, said Umakant Soni, chief executive, Bharat1.AI. 

 

The country has a narrow window of three to five years to get its AI strategy right before global platforms consolidate their dominance, Soni said, adding that India’s stated goal should factor in AI alongside manufacturing and demographic dividend as a growth driver.

 

"We have to build a different ecosystem — there's no two ways about it. And we have to build it fast. Otherwise, we will get drowned in this wave of innovation and capital backing it. Just like we lost in social networking, we could lose here as well," Soni told Business Standard.

 

Bharat1.AI, which counts Nvidia among its partners, is positioning itself as an institutional anchor for that effort. On Sunday, it convened more than 50 leaders from companies including Microsoft, Accel and Neysa at an event in RGA Techpark Bengaluru, to align the industry around what Soni calls a distinctly Indian playbook for AI development.

 

The company, which Soni cofounded earlier this year with Subhashis Banerjee and Sireesh Kupendra, is building what it describes as a humanity-centric AI ecosystem — a phrase Soni uses to mean AI that augments rather than replaces people. Its model draws comparisons with Paris-based Station F, but is focused exclusively on AI and robotics. In a planned second phase, Bharat1.AI intends to build a physical AI city as a live testing environment, generating real-world data for frontier AI research.

 

The warning reflects a broader anxiety in Indian technology circles about whether the country can move quickly enough to avoid becoming a passive consumer of AI platforms developed in the United States and China. Soni argues that India has the raw material — 17 million developers, the second-largest base globally, alongside proven digital public infrastructure such as UPI — but lacks a coordinated strategy to deploy it.

 

The event on Sunday was partly designed to address that gap. Soni said developers across India are not yet building with a unified approach, and that Bharat1 is running build-a-thons to align them around a common framework, with investor backing contingent on adoption of that playbook.

 

Central to Soni's argument is scale. He said the goal is to expand India's developer base to 100 million while enabling as many as 900 million people to use personalised AI agents. This combination could deliver the kind of nonlinear growth India needs to build a $30 trillion economy, he said.

 

Getting it wrong, Soni said, would be costly. He estimated that without the right AI strategy, India's economy could plateau at $8 trillion to $10 trillion by 2047 even as its population grows to around 1.6 billion — a scenario in which the demographic dividend fades rather than compounds. "Getting it wrong means we become consumers," he said.

 

He was blunt about the competitive threat. "If we don't get this right in the next three to five years, we could be subsumed by either China or the US," he said.

 

Soni argues that India's approach to AI should be structurally different from Western or Chinese models. It should be shaped by the country’s specific context instead of being imported wholesale from Silicon Valley or Shenzhen. The context includes 22 official languages, over 6 lakh villages, and a fragmented healthcare system that differs sharply from the environments in which dominant AI models were trained.

 

“India’s linguistic and cultural diversity must shape AI development,” he said, adding that over-reliance on homogenised AI outputs risks eroding that uniqueness. The country remains in the early stages of developing AI systems suited to that context, he added.

 

Soni cautioned against replicating the US model of massive, centralised GPU clusters, arguing it is both energy-intensive and ill-suited to Indian conditions. "Instead of centralising massive GPU clusters, we should focus on distributed intelligence — humans plus AI at the edge," he said.

 

He added that India should look beyond large language models (LLMs) at a broader AI stack. On artificial general intelligence, he was sceptical of the hype: "We are far away from AGI," he said.

 

Soni, an IIT Kanpur alumnus, has spent more than 16 years working in AI, founding what he describes as India's first AI chatbot company in 2009-10 and contributing to national AI policy. 

 

He framed the current moment as a genuine inflection point, one that India has the capacity to seize — but only if it moves deliberately and quickly. “As long as we stay truthful to the path of humanity, we are in safe hands,” he said.