Bengaluru Startup Puts India at Center of Global AI Policy

June 23, 2026

A Bengaluru-based AI startup has emerged as a pivotal voice in global AI policy discussions, showcasing how India’s approach to artificial intelligence—centered on linguistic diversity and sovereignty—is reshaping the global conversation on AI governance.

🇮🇳 Sarvam AI: India’s Sovereign AI Champion

Sarvam AI, founded in August 2023 by Vivek Raghavan and Pratyush Kumar, has become the most prominent symbol of India’s sovereign AI ambitions . The startup, which recently achieved unicorn status with a $1.5 billion valuation** after raising **$234 million in a Series B round led by HCLTech, is building foundational AI models specifically designed for Indian languages and real-world use cases .

At the India AI Impact Summit 2026 in New Delhi, Sarvam unveiled a new generation of large language models, including 30-billion and 105-billion parameter models using a mixture-of-experts architecture, as well as text-to-speech, speech-to-text, and vision models . The company’s Indus model, released as a beta on February 20, is now available across platforms and supports all 22 scheduled Indian languages, allowing seamless switching between English and Indic languages like Hindi, Tamil, or Bengali .

🌍 Global Recognition and Policy Influence

The significance of Sarvam’s work was underscored when Google CEO Sundar Pichai, speaking at the India AI Impact Summit 2026, publicly praised the startup, saying: “You know, even recently, the work Sarvam has done developing local AI models… I just don’t see any impediments to that, and I think it is very, very well positioned” .

This endorsement from one of the world’s most influential tech leaders highlighted a critical shift: India is no longer just a consumer of AI but is becoming a creator of foundational AI technologies that reflect its own linguistic and cultural realities .

Pratyush Kumar, Sarvam’s co-founder, captured the moment: “Today we show we can bring our own AI to a billion Indians” . Co-founder Vivek Raghavan added: “Sovereignty matters much more in AI than building the biggest models” .

🏛️ The Sovereign AI Mandate

The timing of Sarvam’s rise is significant. The US government’s decision to restrict foreign access to Anthropic’s most advanced AI models served as a stark reminder that AI infrastructure controlled by other nations carries inherent policy risk .

For Indian banks, insurers, and government departments that had been building on foreign AI APIs, this was a wake-up call . The Indian government has responded by actively supporting sovereign AI development through the IndiaAI Mission, which provides compute access and funding to select startups .

🤖 What Makes Sarvam’s Models Distinct

Sarvam’s approach to AI is fundamentally different from global competitors in several ways:

Language Inclusivity: The startup’s models are built to be used through voice commands and accessible through 22 Indian languages, giving them a competitive advantage in a country where the vast majority cannot read, write, or type in English .

Cultural Nuance: The models are trained on trillions of Indian data sets, particularly those in Indian languages, enabling them to understand regional accents, slang, and cultural nuances that global models routinely miss .

Superior Performance: Sarvam’s vision model scored 84.3% accuracy on olmOCR-Bench, surpassing Gemini 3 Pro and DeepSeek OCR v2, and even higher on OmniDocBench v1.5 .

Full-Stack Sovereignty: The company describes itself as “Sovereign by design,” with everything developed, deployed, and governed entirely in India .

🌐 India’s Broader AI Policy Influence

Sarvam is not alone. India’s sovereign AI push extends across multiple initiatives:

BharatGen, led by IIT Bombay, has released PARAM-2, a 17-billion-parameter multilingual model supporting all 22 scheduled Indian languages .
Gnani.ai has launched voice AI models that clone voices in 12 Indian languages .
BharatGen has also released SHRUTAM (speech-to-text), SUKTAM (text-to-speech), and PATRAM (India’s first document vision-language model), each built from scratch .

The government’s IndiaAI Mission, with a budget of approximately ₹10,371 crore, is actively supporting the development of indigenous foundational models, providing subsidized compute access to select startups .

🔮 What This Means for India’s AI Future

Sarvam’s rise places India at the center of global AI policy discussions. As one observer noted: “Indian AI startups like Sarvam and BharatGen… are looking to export their AI systems to other developing economies where neither Chinese nor US models are favoured” .

The message is clear: India is moving from being a consumer of global AI to a creator of foundational models for its own unique context—and increasingly, for the world. As Professor Ajay Sood, Principal Scientific Adviser to the Government of India, stated: “India has long believed that science advances fastest when it is open and shared”