Sarvam becomes India’s newest AI unicorn with $234M funding round

A startup that didn’t exist two years ago just joined the billion-dollar club. Sarvam AI, founded in Bengaluru in August 2023, has raised $234M in a funding round that values the company at roughly $1.5B, making it India’s newest AI unicorn.

The round’s headline act: Indian IT giant HCLTech is writing a $150M check, accounting for nearly two-thirds of the total raise. Bessemer Venture Partners led the round, with Nvidia and Amazon also participating.

What Sarvam actually builds

Sarvam AI isn’t trying to out-GPT OpenAI. Instead, the company is building what it calls a “full-stack sovereign AI platform” designed from the ground up for India’s linguistic complexity. India has 22 officially recognized languages, and most Western AI models treat them as an afterthought, fine-tuned translations bolted onto English-first architectures.

In February 2026, the company released two large language models: Sarvam-30B and Sarvam-105B. Both use a mixture-of-experts (MoE) architecture. In English: instead of one massive neural network trying to do everything, MoE models route different tasks to specialized sub-networks, making them more efficient without sacrificing capability.

Both models were trained from scratch on Indic-language datasets. The company’s vision aligns with India’s broader IndiaAI Mission, a government-backed initiative pushing for technological self-reliance. Co-founders Vivek Raghavan and Pratyush Kumar have positioned Sarvam squarely within that national priority.

The funding trajectory tells a story

Before this round, Sarvam had raised approximately $41M across seed and Series A stages. Investors in those earlier rounds included Lightspeed Venture Partners, Peak XV Partners (the former Sequoia India), and Khosla Ventures.

HCLTech’s $150M commitment deserves its own analysis. HCLTech is one of India’s largest IT services companies, with deep enterprise relationships across banking, healthcare, manufacturing, and government. The investment signals that HCLTech likely plans to integrate Sarvam’s language models into its own service offerings.

What this means for investors

The competitive landscape is worth watching. Sarvam isn’t the only Indian AI startup chasing this opportunity. Krutrim, founded by Ola’s Bhavish Aggarwal, also targets Indian-language AI and reached unicorn status in early 2024. Sarvam’s technical approach—building models from scratch on native datasets rather than fine-tuning existing architectures—represents a more ambitious strategy.

For crypto-native investors, the sovereign AI narrative has indirect but meaningful implications. Decentralized AI projects, many of which trade on the premise that AI infrastructure should not be concentrated in a few corporate hands, share philosophical DNA with the sovereign AI movement. Sarvam itself has no blockchain involvement.

The risk factors are real. Building large language models from scratch is enormously capital-intensive, and $234M gets consumed quickly when training 105-billion-parameter models. The research notes the total round size may be $300–350M according to some reports, against the $234M figure cited elsewhere.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.