India is the second-largest market for OpenAI and Anthropic after the United States. Yet the country still has few companies trying to build frontier models at home, largely because compute is expensive and late-stage AI capital has been scarce.
Sarvam has now filled part of that gap. On June 15, the Indian AI company announced a $234 million first close of a planned $300 million Series B. The round values Sarvam at $1.5 billion and turns it into a unicorn. HCLTech led the financing with $150 million for a stake of more than 10%, while Bessemer joined and existing backers Khosla Ventures and Peak XV followed on.
The jump is unusually sharp: before this round, Sarvam had raised only $41 million across its seed and Series A financings.
What Sarvam Is Building
Sarvam describes itself as a full-stack sovereign AI company for India. In its framing, sovereign AI means keeping the value loop of data, models and agents inside a country or an enterprise instead of depending entirely on foreign platforms.
The company is working from infrastructure upward: training and inference systems, frontier text and multimodal models, and products for enterprises, developers and government users. Its priority sectors include banking, insurance, public services and defense.
- 2 million conversational AI interactions a day
- 10 million API calls a day on its inference platform
- 500,000 hours of audio transcribed each month
- 35 million pages digitized by its document AI
- 17 million farmers' records collected for India's agriculture ministry
- Voice outreach for a major insurer covering 45 million policyholders
- An agent platform for a fintech sales force of 350,000 people
"Our ambition is to diffuse this technology widely in India, creating significant value across sectors for citizens, small businesses, enterprises, and state and central governments."
Where The Money Goes
Sarvam says the new capital will go into two areas: training its next generation of frontier models, with an emphasis on agents, coding and cybersecurity, and expanding compute so its closely deployed AI stack can reach more industries.
Those three model directions are among the most valuable and contested in global AI. An Indian company openly funding that roadmap is itself a signal that the local market wants more than API resale.
Why The Timing Matters
The timing is hard to ignore. Days earlier, Anthropic cut off access to its Fable 5 and Mythos 5 models under US export-control pressure, including for some overseas users. For governments and companies that have built critical workflows on American models, the lesson was blunt: access can change because of policy, not just product quality.
That turns sovereign AI from a slogan into a budget line. India is a huge AI market, but long-term dependence on US models is now seen as a strategic risk. Sarvam's round is a bet that local model capacity and domestic data control will be worth paying for.
HCLTech's $150 million commitment is also more than a financial wager. As a strategic investor, it can bring Sarvam into the large enterprise accounts it already serves. The harder question is whether Sarvam's next models can compete with open Chinese models and closed US leaders. Either way, India now has a serious domestic contender.
Sources: CocoLoop, TechCrunch report on Sarvam's $234 million funding round led by HCLTech, and HCLTech's announcement of the first close of Sarvam's $300 million Series B at a $1.5 billion valuation.