India’s $2 Billion AI Bet: Yotta Builds the Subcontinent’s GPU Empire

In a server room on the outskirts of Mumbai, rows of Nvidia H100 chips hum with the quiet intensity of a technological revolution. This is Yotta Data Services’ flagship facility, and it’s rapidly becoming the beating heart of India’s artificial intelligence ambitions.

“We own 60% to 70% of India’s GPU capacity. Demand is exceeding supply as domestic AI models prepare to scale.” — Sunil Gupta, CEO of Yotta Data Services

A $2 Billion Infrastructure Play

Yotta Data Services is betting big on India’s AI future. The Mumbai-based company is building a $2 billion artificial intelligence hub powered by Nvidia’s most advanced chips. It’s a gamble that could determine whether India becomes a major player in the global AI race or remains a consumer of technologies built elsewhere.

The timing is critical. India currently trails the United States and China in developing native AI foundational models. The country lacks the massive domestic infrastructure that underpins AI development in those superpowers. But that gap is beginning to close, and Yotta intends to be the bridge.

Most Indian AI models launched at last week’s India AI Summit were trained on Nvidia GPUs hosted in Yotta’s facilities, according to Gupta. That includes Sarvam AI’s Indus chatbot, one of the first homegrown large language models designed specifically for Indian languages and contexts.

The Global Rush for Indian Users

While domestic startups build on Yotta’s infrastructure, global AI giants are racing to capture India’s 1.4 billion potential users. The strategy is familiar: offer services at low or no cost, build market share, and monetize later.

OpenAI has made India a centerpiece of its international expansion, launching “OpenAI for India” and becoming the first customer of Tata Consultancy Services’ data center business with a 100 MW capacity commitment. “We’re working together to build the infrastructure, skills, and local partnerships needed to build AI with India, for India, and in India,” CEO Sam Altman said in a statement.

Google has committed $15 billion to build a data center hub in southern India. Microsoft plans to invest $17.5 billion to expand its data center footprint. Combined with domestic investments, companies announced plans to pour $277 billion into Indian AI infrastructure over the next five to seven years, according to Nomura.

“As the Indian user base of leading global AI companies expands, they will require local data centers and GPU capacity. We’re positioning Yotta to be that provider.” — Sunil Gupta

The Supply Crunch

Yotta’s dominance comes with challenges. GPU demand in India is already exceeding supply, creating bottlenecks for companies trying to train and deploy AI models. Sarvam AI’s co-founder Pratyush Kumar acknowledged the constraint when launching Indus: “We’re gradually rolling out on limited compute capacity, so you may hit a waitlist at first.”

The shortage isn’t unique to India. A global squeeze on memory chips sparked by the AI boom has sent shockwaves through the technology sector. CNN reported this week that the shortage has dealt a “tsunami-like shock” to the smartphone industry as manufacturers compete with AI data centers for limited semiconductor supply.

For Yotta, the constraint represents both a challenge and an opportunity. The company began sourcing Nvidia GPUs in 2023 and has steadily built inventory. Now it’s planning a $1.2 billion to $1.5 billion pre-IPO funding round to purchase additional chips and expand capacity. Gupta aims to take the company public within the next 12 months.

India’s Data Center Ambitions

The numbers paint a picture of rapid transformation. India had 1.93 gigawatts of total data center capacity in 2025. By 2028, that figure is projected to nearly double to 4 gigawatts, according to Nomura.

The majority of new investment will flow into AI-optimized facilities. Domestic firms and U.S. technology companies are leading hyperscale buildouts, positioning India as a key American technology partner in a region where China has traditionally dominated.

Whether Yotta can maintain its market-leading position as competition intensifies remains an open question. The company has first-mover advantage and deep relationships with Nvidia. But global cloud providers with deeper pockets are circling, and India’s regulatory environment for data localization adds complexity to every expansion decision.

For now, the GPUs keep humming in that Mumbai server room. And Sunil Gupta is betting that in the global AI race, controlling the infrastructure means controlling the future.

This article was reported by the ArtificialDaily editorial team. For more information, visit CNBC and Reuters.