India Bets Big on Sovereign AI With Sarvam's $350M Raise

While the U.S. and China dominate AI headlines, India is quietly assembling the infrastructure for its own AI ecosystem — and the latest signal is impossible to ignore. Sarvam AI, a Bangalore-based startup building voice-first AI systems for India's 22 official languages, is closing a $300 million to $350 million funding round at a valuation of $1.5 billion.

Bessemer Venture Partners is expected to lead the round, with Nvidia, Amazon, and Prosperity7 Ventures also participating. The deal marks one of the largest AI funding rounds outside the U.S. and China and positions Sarvam as the clear frontrunner in India's emerging sovereign AI race.

Why Sovereign AI Matters in India

India has over 1.4 billion people, but a large share of them don't primarily interact in English. That's a fundamental problem for AI systems trained predominantly on English-language data. Sarvam AI was founded in 2023 by AI researchers Vivek Raghavan and Pratyush Kumar specifically to address this gap — building models that handle Hindi, Tamil, Bengali, Telugu, and 18 other Indian languages natively.

The company's 105-billion-parameter model, launched in February 2026, outperforms global competitors on what's called "Indic code-switching" — the ability to handle mixed Hindi-English text natively, which is how hundreds of millions of Indians actually communicate. If you've ever texted in Hinglish, you understand why this matters.

The Government Is All In

India's central government isn't just watching from the sidelines. Through its IndiaAI Mission, the government has transferred approximately $11 million in GPU subsidies to Sarvam — specifically, 4,096 Nvidia H100 SXM GPUs provisioned through Yotta Data Services. That's the largest single allocation the IndiaAI Mission has made, out of a total corpus backed by over $1.2 billion in the 2026-27 Union Budget.

The government is also directly funding 12 organizations to build sovereign AI foundation models, including Gnani AI, Fractal Analytics, and IIT Bombay's BharatGen consortium. The BharatGen project alone received four times the allocation of the next-highest funded initiative at roughly $120 million.

Microsoft's $17.5 Billion Commitment

The private sector is matching government ambition dollar for dollar. Microsoft's $17.5 billion India investment — its largest in Asia — is deploying through 2029. A flagship hyperscale data center in Hyderabad launches this summer with three availability zones, providing the kind of compute infrastructure that AI development demands.

Microsoft has also doubled its pledge to train 20 million Indians in AI skills by 2030. The combination of compute infrastructure, talent development, and government backing creates the foundation for an AI ecosystem that could eventually challenge dominant U.S. and Chinese players in emerging markets.

The Competitive Landscape

Sarvam isn't the only Indian AI company attracting capital. Krutrim, backed by Ola founder Bhavish Aggarwal, achieved unicorn status with $50 million in equity plus $230 million in committed financing. Aggregate venture capital for India's top AI companies now exceeds $2.9 billion, and the country hosts approximately 1,700 AI companies.

But Sarvam's edge is specific and defensible: multilingual, voice-first AI for a market where most users don't type in English. In a world where frontier models from OpenAI, Anthropic, and Google optimize primarily for English, there's a real opening for companies that solve language diversity at the model level rather than through translation layers.

What This Means for the Global AI Race

India's AI push matters beyond its borders. As the world's most populous country and fifth-largest economy, India represents an enormous market that global AI companies have struggled to serve effectively. If Sarvam and its peers succeed in building models that genuinely work across India's linguistic diversity, they'll have a playbook that applies to Southeast Asia, Africa, the Middle East, and other multilingual markets.

The sovereign AI movement is also about strategic independence. Countries increasingly view AI capability as critical infrastructure — on par with semiconductors, energy, and telecommunications. India's bet is that it can't afford to depend entirely on American or Chinese AI systems for its digital future.

What to Watch

Sarvam's round is expected to close within days. Once it does, watch for how the company deploys capital. The key metrics will be model performance across Indian languages, enterprise adoption within India's massive IT services sector, and government contract wins as the IndiaAI Mission scales up.

The Bottom Line

India's AI story is no longer about cost arbitrage or outsourcing. With $350 million flowing into Sarvam AI, $17.5 billion in Microsoft infrastructure commitments, and direct government funding for sovereign models, India is making a serious bid to build AI that serves its own population first. In a global AI race dominated by two superpowers, that's a bet worth watching closely.