Sarvam AI Raises ₹230 Crore Series A to Scale Indic Language Models Across 22 Languages
Amit Yadav
Sarvam AI has raised ₹230 crore to expand its 22-language Indic LLM suite, targeting the 900 million Indians underserved by English-only AI products.
Bengaluru-based Sarvam AI has closed a ₹230 crore ($27.5 million) Series A round led by Peak XV Partners, with participation from Lightspeed India and existing angel investors. The capital will be deployed to expand its open-weight language models — currently supporting 22 Indian languages — and build a speech-to-speech AI platform for enterprise clients.
Why Indic LLMs Matter
Over 900 million Indians are not fluent in English. Despite this, nearly all major commercial AI products — from customer support chatbots to coding assistants — are designed primarily for English speakers. Sarvam's thesis is that the next wave of AI adoption in India must be voice-first and vernacular-first.
The company's flagship model, Sarvam-2B, is a 2-billion parameter model trained on curated datasets across Hindi, Tamil, Telugu, Kannada, Malayalam, Marathi, Bengali, Gujarati, and 14 other languages. The model has outperformed GPT-3.5 on several Indic NLP benchmarks and is available under an Apache 2.0 licence on Hugging Face.
Enterprise Traction
Sarvam already counts HDFC Bank, Practo, and the Rajasthan government among its customers. Its voice AI platform handles over 10 million calls per month, automating loan collections, patient appointment reminders, and citizen grievance redressal in regional languages.
"We are not building GPT for India. We are building the AI layer that makes India's existing digital infrastructure actually accessible to the next billion users," said co-founder Vivek Raghavan.