Sarvam AI Just Became a Unicorn, Here’s the $234 Million Story Behind It

Summary

• Sarvam AI raised $234 million of a $300 million Series B, hitting a $1.5 billion valuation and unicorn status.
• HCLTech led with $150 million for a 10.46% stake; Bessemer, Khosla Ventures, and Peak XV Partners also joined.
• Funds will fund a next frontier model and GPU access, with the remaining $66 million still to close.

Sarvam AI’s Series B funding round just made it India’s newest AI unicorn, and the way the deal is structured says as much as the number does. On June 15, the Bengaluru startup announced it had raised $234 million in the first close of a $300 million round, taking its valuation to $1.5 billion. What’s different here is who wrote the biggest check. It wasn’t a Silicon Valley fund. It was HCLTech, an Indian IT services giant putting real money behind a domestic AI lab for the first time at this scale.

What Actually Happened

Sarvam closed $234 million of a planned $300 million Series B, with HCLTech leading as strategic investor at $150 million, enough to take a 10.46% stake worth roughly ₹1,427 crore. Bessemer Venture Partners co-led the round. Existing backers Khosla Ventures and Peak XV Partners returned for another round, while Lightspeed Venture Partners, an early investor, sat this one out.

The company hasn’t said when it expects to close the remaining $66 million, though there’s been speculation about Nvidia joining a later tranche. Co-founder Vivek Raghavan declined to confirm that directly, telling Business Standard only that people should wait for the second closing.

From $41 Million to $1.5 Billion

Sarvam’s funding history is a useful reminder of how fast Indian AI valuations have moved. The startup, founded in 2023 by Vivek Raghavan and Pratyush Kumar, both veterans of the IIT Madras language-AI initiative AI4Bharat, raised $41 million across its seed and Series A rounds back in December 2023. That round, led by Lightspeed, valued the company at roughly $110 million. Eighteen months and one product cycle later, Sarvam is worth more than ten times that.

The jump tracks two things: a string of model releases and a string of enterprise deployments that turned Sarvam from a research project into something companies actually pay to use.

What Sarvam Actually Builds

Sarvam describes itself as a full-stack sovereign AI company, meaning it builds everything from foundation models to the enterprise applications running on top of them, all trained and hosted in India. Earlier this year it released two open-source models built entirely from Indian infrastructure and data: Sarvam 105B, a reasoning model the company says matches larger international models on knowledge and agentic benchmarks, and Sarvam 30B, a lighter model designed to run on consumer-grade hardware.

The usage numbers behind those models are what likely got investors’ attention. Sarvam’s conversational AI platform now handles more than 2 million interactions a day, roughly double its volume from two months ago. Its inference platform processes around 10 million API calls daily, triple what it was three months back. Its speech models transcribe over 500,000 hours of audio every month, and its document AI tool, Sarvam Vision, has digitised more than 35 million pages of records, from insurance forms to old land documents.

Those tools are already running inside real institutions. A multilingual voice agent built for the Ministry of Agriculture and Farmers Welfare collected data from 17 million farmers. A nationwide voice campaign for a major insurer supported policy renewals for 45 million policyholders, and Sarvam has separately partnered with SBI Life Insurance on customer engagement. On the enterprise side, a large fintech company is using Sarvam’s agentic platform to support a sales force of more than 350,000 people.

Read More: How Drools Became the First Pet Food Unicorn of India in 2025

Why HCLTech Specifically

Indian conglomerates have backed startups before, usually through small, symbolic checks from corporate venture arms. A $150 million lead investment from an IT services company is a different kind of bet. HCLTech CEO C Vijayakumar framed it as a step toward building India’s own competitive AI ecosystem, and the stated plan is to pair Sarvam’s models with HCLTech’s enterprise client relationships, engineering workforce, and existing software business to sell AI products directly to banks, insurers, and government bodies that already work with HCLTech.

Bessemer’s Pankaj Mitra made a related point in the company’s announcement: India’s AI ambitions need more than good foundation models. They need the full stack, infrastructure, data, applications, and deployment, built domestically. That’s the bet Sarvam’s investor list is making collectively.

Fresh capital from the round is earmarked for training Sarvam’s next frontier model, focused on agentic AI, coding, and cybersecurity use cases, along with securing more GPU access. Raghavan has said anywhere from 30% to 50% of the funds will likely go toward compute procurement alone, which gives a sense of how capital-intensive frontier AI work still is, even for a company already generating real revenue.

Read More: India’s First Homegrown Sovereign AI Model, Sarvam AI To Build

Startup INDIAX Take

What makes this round worth watching isn’t just the unicorn label, it’s who’s funding it. Most of India’s AI capital so far has come from the same global funds that back Silicon Valley startups. HCLTech leading at $150 million signals that domestic capital is now willing to underwrite frontier AI research, not just buy services from it. For founders building in deep tech or infrastructure-heavy categories, that’s a meaningful shift in who’s available to write the big checks. For India’s broader AI ambitions, it suggests the “sovereign AI” pitch has moved from policy talking point to something investors are pricing into real valuations.

Why This Matters

For Indian AI founders, Sarvam’s round is proof that building foundation models domestically, rather than wrapping a foreign model in a local interface, can attract serious capital and serious enterprise customers. For investors, it’s a signal that AI infrastructure plays in India are no longer purely speculative; Sarvam’s usage numbers show real institutional adoption, not just demo traction. For enterprises and government bodies, the HCLTech partnership specifically points toward sovereign AI products becoming commercially available at scale, rather than staying confined to pilot programs. And for consumers, it means more AI tools built for Indian languages and use cases, deployed through institutions people already interact with, from insurers to government agriculture programs.

The Bigger Picture

Sarvam’s round lands at a moment when sovereign AI has become a serious global theme, not just an Indian one. Governments from Europe to the Gulf are funding domestic AI infrastructure as a hedge against depending entirely on US or Chinese frontier labs. India’s own push, through the IndiaAI Mission and its planned compute infrastructure, has already backed multiple homegrown model efforts, and Sarvam was among the first startups selected under that program.

What’s notable is the capital source. Indian IT services firms have deep enterprise relationships but historically thin AI research capability. Foreign-funded Indian AI startups have research talent but limited enterprise distribution inside India’s regulated sectors. Sarvam’s deal effectively merges the two, and if it works, it could become a template other Indian conglomerates look to repeat with other AI startups over the next year.

Read More: Krutrim’s Bold Leap: Unveiling Its Agentic AI Assistant Kruti in 2025

Got thoughts on whether Indian conglomerates backing AI startups is the playbook other founders should be chasing? Drop them in the comments, and check out more of Startup INDIAX’s funding coverage for the deals shaping India’s AI race this year.

FAQs