Sarvam AI Cofounder Vivek Raghavan Bats For ‘Sovereign AI’

Amid the ongoing global AI race, India must focus on building a ‘sovereign AI’ ecosystem to ensure autonomy over its data, Sarvam AI cofounder Vivek Raghavan has said. 

With GenAI increasingly penetrating diverse sectors, there is an urgent need for sovereign AI, Raghavan said today at Inc42’s ‘The GenAI Summit 2025’.

“… We have always wanted to think of the concept of sovereign AI … how do you control your own data, your own models and your own compute?” Sarvam AI’s cofounder said.

Founded in 2023 by Raghavan and Pratyush Kumar, Sarvam AI claims to have launched India’s first open source foundational small language model, Sarvam 2B, with a focus on Indic languages. The startup claims that Sarvam 2B is trained with 4 Tn tokens and performs better than Meta’s Llama in Indian languages. 

Notably, Sarvam AI is among the startups that have submitted proposals to the Centre to build India’s own foundational large language model (LLM).

In December 2023, the Bengaluru-based startup raised $41 Mn (around INR 342 Cr) in its Series A funding round led by Lightspeed Venture Partners, in participation with Peak XV Partners and Khosla Ventures.

Raghavan’s comments come amid rising calls in the country for a sovereign AI. Last month, Tata Sons chairman N Chandrasekharan said it is imperative for India to focus on sovereign AI capabilities to avoid “digital colonialism”.

The US currently has a near hegemony in the global AI market with its deep-pocketed Silicon Valley titans such as OpenAI, Meta and Google. However, China is emerging as a threat with its DeepSeek, an AI model which is said to have been built at a fraction of cost and using much far less computing power than its US counterparts.

On the other hand, India is behind its peers and is yet to develop an LLM that can compete with the likes of OpenAI’s GPT-4o, Anthropic’s Claude Sonnet 3.5, DeepSeek-r1, and others. 

Bracing up for the AI challenge, India is looking to build a domestic LLM under the INR 10,370 Cr IndiaAI Mission.

The Centre has received more than 180 proposals to build foundational AI models, including applications from local AI startups such as Sarvam AI, CoRover and Ola’s Krutrim. The government has also selected 10 companies that will supply 18,693 graphics processing units or GPUs — high end chips needed to develop machine learning tools that can go into developing a foundational model. 

India is home to more than 200 GenAI startups that raised more than $1.2 Bn in funding between 2020 and 2024. 

Overall, the Indian AI ecosystem is projected to become a $17 Bn market opportunity by 2030.