Sarvam AI Raises $234M to Train Frontier AI Models with 30B-105B Parameters

While many AI labs chase ever-larger models, Sarvam AI is proving that parameter count isn't everything—context matters. With open-source releases of 30B and 105B models earlier this year and an ambitious roadmap into agentic, coding, and cybersecurity AI, the $234 million Series B will fuel frontier research on sovereign infrastructure. For AI practitioners, Sarvam offers a blueprint for building mission-critical models that speak the local language—literally—and operate within constrained compute budgets.

Sarvam AI, the Bengaluru-based sovereign AI startup, has cemented its position as India's newest unicorn, raising $234 million in the first close of its Series B round at a post-money valuation of $1.5 billion. The round was led by HCLTech, the IT services arm of the HCL Group, which invested $150 million for a 10.46% stake, making it the lead strategic investor. Bessemer Venture Partners also joined the round, while existing backers Khosla Ventures and Peak XV Partners doubled down. Sarvam is targeting a total Series B haul of $300 million.

“

Sarvam AI, the Bengaluru-based sovereign AI startup, has cemented its position as India's newest unicorn, raising $234 million in the first close of its Series B round at a post-money valuation of $1.5 billion.

The funding marks a significant leap from the $41 million the company raised across its seed and Series A rounds, and it follows the launch earlier this year of Sarvam's open-source language models at scales of 30 billion and 105 billion parameters. This capital infusion reflects a broader sovereign AI push by India, where both public and private entities seek to develop homegrown alternatives to Western large language models and cloud infrastructure. Sarvam sits at the intersection of this shift, offering a full-stack platform spanning model training, inference infrastructure, and enterprise applications — all tailored to Indian languages and use cases.

Sarvam's commercial traction is already substantial. Its multilingual voice agents have collected data from 17 million farmers for India's Ministry of Agriculture and Farmers Welfare, while a nationwide voice campaign for a leading insurer supported policy renewals for 45 million policyholders. The platform now handles over 10 million API calls daily and is digitizing 35 million pages of handwritten and Indian-language records through its Sarvam Vision document AI platform. These numbers underscore the startup's ability to deploy at India's massive scale, a key selling point for both investors and government clients.

HCLTech's involvement is more than a financial bet. The partnership will meld Sarvam's AI research with HCLTech's enterprise expertise, global client relationships, and engineering workforce to co-create AI products for sectors including banking, insurance, government services, and defense. HCLTech CEO C Vijayakumar described the investment as a step toward building India's trusted, globally competitive AI ecosystem. For Sarvam, the alliance provides a deep-pocketed commercialization channel and access to a vast enterprise customer base.