India has a new AI unicorn. Sarvam, a Bengaluru-based startup, raised $234 million at a $1.5 billion valuation, the company announced Monday. According to TechCrunch AI, the round is led by HCLTech, the IT arm of Indian conglomerate HCL Group, which is putting in $150 million as lead strategic investor. Bessemer Venture Partners joined too, alongside existing backers Khosla Ventures and Peak XV Partners. Sarvam wants to close the full Series B at $300 million.
What stands out here is the jump in scale. Sarvam raised just $41 million across its seed and Series A rounds more than two years ago. This single round is more than five times that, and it lands shortly after the startup shipped open source models at 30 billion and 105 billion parameters earlier this year.
Why this matters
This is significant because it’s a bet on sovereign AI. Governments and companies increasingly want control over the models and computing infrastructure they depend on, rather than renting access from a handful of foreign providers. Sarvam is one of the few Indian firms trying to build the whole stack: the models, the inference infrastructure, and the enterprise apps on top.
The timing isn’t an accident. The sovereignty debate got louder last week when Anthropic disabled access to its latest models, Fable 5 and Mythos 5, after the U.S. government ordered the company to suspend their use by any foreign national on national security grounds. As TechCrunch AI notes, that move showed just how concentrated access to frontier AI really is. For a country building national-scale AI services, that’s a risk worth $234 million to hedge against.
The India context
India is a giant AI consumer but a small producer of frontier models. Both OpenAI and Anthropic have called it their second-largest market after the U.S., powered by a huge base of developers, enterprises, and consumers. Yet high compute costs and limited capital have kept Indian startups out of the frontier race against well-funded rivals in the U.S. and China. Sarvam is one of the few trying to change that with homegrown foundation models built for Indian languages and use cases.
The HCLTech tie-up is the commercial engine. The plan is to pair Sarvam’s models with HCLTech’s enterprise client relationships, engineering workforce, and software assets to sell AI products to businesses and governments. That gives Sarvam a distribution channel most model labs would envy.
The numbers behind the bet
Sarvam isn’t a research project. The company shared real deployment figures:
- Its conversational AI platform handles more than 2 million interactions a day.
- Its inference platform processes roughly 10 million API calls daily.
- Its speech models transcribe over 500,000 hours of audio each month.
- Its document AI is digitizing more than 35 million pages of records.
The real-world use cases are striking. Sarvam’s multilingual voice agents collected data from 17 million farmers for India’s Ministry of Agriculture and Farmers Welfare. A nationwide voice campaign for a major insurer supported policy renewals for 45 million policyholders. And a large fintech is using Sarvam’s agentic platform to back a sales force of more than 350,000 people. Its products already reach banking, insurance, government services, and defense.
What comes next
Sarvam says the fresh capital will fund its next generation of models aimed at agentic, coding, and cybersecurity uses, plus more computing infrastructure to scale deployments. The startup was founded by Vivek Raghavan and Pratyush Kumar, both veterans of AI4Bharat, the Indian-language AI research effort at IIT Madras backed by Nandan Nilekani.
“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,” Raghavan said.
Watch for whether Sarvam can turn deployment scale into the kind of model quality that competes with overseas frontier labs. If it can, it becomes a template for how mid-sized economies build AI they actually control. You can find the full details in the original report from TechCrunch AI.