India's Largest AI Model Round
On 15 June 2026, Bengaluru-based Sarvam AI announced a 234-million-dollar Series B first close at a 1.5-billion-dollar valuation — becoming the latest, and arguably most strategically significant, AI unicorn in India. The round was led by a 150-million-dollar investment from HCLTech, one of India's largest IT services companies, which acquires a stake of more than 10 per cent in the process. Bessemer Venture Partners participated alongside existing investors Khosla Ventures and Peak XV Partners. Sarvam has stated a target of 300 million dollars for the total round, placing the first close at 78 per cent of that target.
The round represents the largest single raise by an India-based AI model company. It also carries explicit sovereign-AI intent: Sarvam was built from the start to develop full-stack AI — large language models, speech systems, document AI — that runs natively on Indian languages and is designed to remain under Indian control.
What Sarvam Builds
Sarvam AI was founded in 2023 by Dr. Vivek Raghavan and Dr. Pratyush Kumar. Dr. Raghavan played a formative role in the Aadhaar digital identity programme, the world's largest biometric identity system with over 1.3 billion enrolled individuals. Dr. Pratyush Kumar trained at IIT Bombay and ETH Zurich and conducted AI research at Microsoft Research and IBM Research before co-founding Sarvam.
The company builds large language models fine-tuned on Indian languages, speech recognition systems capable of handling the phonological diversity of Indic scripts, vision-language models, and enterprise-grade document AI. Unlike global AI companies that treat Indian languages as a downstream localisation task, Sarvam trains its models specifically for the linguistic structures of Hindi, Tamil, Telugu, Marathi, Kannada, and other Indic languages — 22 scheduled languages that together cover more than a billion speakers.
The company's flagship enterprise deployment demonstrates scale: a major Indian fintech company uses Sarvam's agentic AI platform to support a sales force of more than 350,000 people — handling queries, workflows, and transactions in the agent's language rather than requiring English-first interaction.
HCLTech's Strategic Investment
HCLTech's 150-million-dollar investment is larger than a standard minority stake — it is a strategic alignment between India's IT services industry and India's most advanced domestic AI model company. For HCLTech, which competes with Infosys and Tata Consultancy Services for AI-led services contracts, direct access to a sovereign-model-capable partner gives it a differentiated proposition in tenders where data localisation or Indian language coverage is a requirement.
For Sarvam, the HCLTech capital and distribution network accelerates the path from research-grade model to enterprise deployment at a scale that pure venture capital timelines rarely allow.
Sovereign AI and Why It Matters
Dr. Raghavan has been explicit about the geopolitical dimension of Sarvam's mission. In a February 2026 public statement, he warned that large global AI companies are providing free-tier model access as a mechanism to capture long-term infrastructure control — and called on the Indian government to act as the first buyer of domestically built AI. The IndiaAI Mission, a government-backed programme to develop and deploy sovereign AI infrastructure, is the policy context in which Sarvam is operating.
Sarvam is not alone in this position: the same argument underpins AI investment policy across the European Union and Southeast Asia. But India's scale — a market of 1.4 billion users who predominantly prefer their native language over English — makes the opportunity for a domestically built model company uniquely large.
The Next Frontier: 15 Trillion Tokens and Agentic AI
The fresh capital will fund the next phase of model development. Sarvam has disclosed plans for a next-generation frontier model trained on 15 trillion tokens with specific focus on Indic fine-tuning — a scale that would make it the largest India-origin LLM training run on record. Target domains include agentic AI for complex multi-step enterprise workflows, coding assistance in Indian enterprise contexts, and cybersecurity applications.
Microsoft has also confirmed collaboration with Sarvam to deploy its voice-based generative AI models on Azure, extending the company's reach into the cloud infrastructure that most Indian enterprises already use.
What This Means for Indian Technology Builders
For software teams building products for Indian users, Sarvam's unicorn round has a direct practical implication: the infrastructure for native Indian-language AI is now sufficiently funded and mature to build on. Applications that have historically required English-first design with patchy regional language support can now be architected around voice-first, Indic-language-native foundations.
For enterprise development teams, the HCLTech partnership signals that Sarvam's models will increasingly be available through existing services relationships and cloud channels — reducing the integration friction that has historically made sovereign AI options less accessible than global alternatives. Teams building customer-facing applications for Tier 2 and Tier 3 Indian markets, where English is rarely the primary language, should evaluate Sarvam's API as a first-class option rather than a regional add-on.
The Bottom Line
Sarvam AI's 234-million-dollar Series B at a 1.5-billion-dollar valuation, announced 15 June 2026 and led by HCLTech's 150-million-dollar investment, is the largest raise by an India-origin AI model company. It funds a next-generation model trained on 15 trillion tokens for Indian languages, agentic AI enterprise products, and coding and cybersecurity capabilities. For Indian technology teams building products for a billion-plus native-language audience, Sarvam's growth represents the arrival of a sovereign AI infrastructure option that is now enterprise-scale, nationally backed, and sufficiently funded to build on with confidence.
Frequently Asked Questions
Who are the founders of Sarvam AI and what does the company build?+
Sarvam AI was co-founded in 2023 by Dr. Vivek Raghavan and Dr. Pratyush Kumar. Dr. Raghavan played a key role in the Aadhaar biometric identity programme — the world's largest digital identity system. Dr. Pratyush Kumar holds a PhD from ETH Zurich and conducted AI research at Microsoft Research and IBM Research. The company builds large language models, speech recognition systems, document AI, and vision-language models specifically trained for India's 22 official languages including Hindi, Tamil, Telugu, Marathi, and Kannada.
What are the terms of Sarvam AI's Series B funding round?+
Sarvam AI announced a 234-million-dollar Series B first close on 15 June 2026 at a 1.5-billion-dollar valuation, achieving unicorn status. HCLTech led the round with a 150-million-dollar strategic investment, acquiring more than 10 per cent of the company. Bessemer Venture Partners also participated, alongside existing investors Khosla Ventures and Peak XV Partners. The company is targeting a total raise of 300 million dollars for the full round.
What is sovereign AI and why is it central to Sarvam's mission?+
Sovereign AI refers to the development of artificial intelligence infrastructure — models, training data, and compute — that remains under national or organisational control rather than relying on foreign technology providers. Sarvam co-founder Vivek Raghavan has warned that large global AI companies offer free-tier model access as a strategy to capture long-term infrastructure control. Sarvam builds Indian-language-native models as an alternative that keeps AI capability and data within India, operating under the policy umbrella of the government-backed IndiaAI Mission.
What will Sarvam build with the new funding?+
The 234-million-dollar raise will fund a next-generation frontier model trained on 15 trillion tokens with specific Indic language fine-tuning — the largest India-origin LLM training run on record if completed. Target application areas include agentic AI for enterprise workflows, coding assistance tailored to Indian enterprise contexts, and cybersecurity applications. Microsoft has also committed to deploying Sarvam's voice-based AI models on Azure, extending the company's enterprise reach through existing cloud channels.
Written by
TechPillow Team
Sharing insights on technology, product development, and the Indian tech ecosystem.