Launch · AI & Deep Tech

Sarvam AI launches Indic-focused open-source models

By Startup Enthusiast · 13 March 2026

What they do

Sarvam AI builds sovereign AI architecture with 30B and 105B open-source mixture-of-experts models

It released models with custom tokenisers for Indian scripts like Odia and Manipuri

The company trails global AI giants in long-context capabilities by 18 months

Sarvam AI builds full-stack sovereign AI architecture with 30B and 105B open-source mixture-of-experts models.
It uses a custom tokeniser designed for scripts like Odia and Manipuri, requiring less compute for complex Indian dialects.
The models are deployed at Aphelion Labs for a voice-first rural banking assistant.
Sarvam AI's comparative advantage is Indic performance beyond Hindi and English.
Sarvam AI trails global AI giants in long-context and action model capabilities by nearly 18 months.
The company is offering credits and hosting founder sessions to tap into the homegrown startup ecosystem.

Sarvam AI builds full-stack sovereign AI architecture with 30B and 105B open-source mixture-of-experts models.
It uses a custom tokeniser designed for scripts like Odia and Manipuri, requiring less compute for complex Indian dialects.
The models are deployed at Aphelion Labs for a voice-first rural banking assistant.
Its comparative advantage is Indic performance beyond Hindi and English.
It faces adoption friction: absence of GGUF formats for local execution, lack of day-one support for vLLM, and broken chat templates.
Sarvam AI trails global AI giants in long-context and action model capabilities by nearly 18 months.

Sarvam AI is offering credits and hosting founder sessions to tap into the homegrown startup ecosystem.

Product: Foundation models and agentic AI systems with voice capabilities and enterprise applications
Customers: Enterprise customers across finance, aviation, telecom and government sectors
How it makes money: SaaS and licensing model for AI models and enterprise AI applications
What sets it apart: India-built sovereign AI stack with focus on cost-effective deployment and localization