Sarvam AI has unveiled an artificial-intelligence model that is more suitable for India than ChatGate and Cloud.
The Bengaluru-based AI startup announced the two models at the India AI Impact Summit 2026 in New Delhi – a showcase of Prime Minister Narendra Modi’s efforts to make his country a leading player in the emerging technology. Sarvam’s models are built to be used via voice command and are accessible through 22 Indian languages, something the company says will be a competitive advantage in a country of 1.45 billion where the vast majority cannot read, write or type in English.
“Today we have shown that we can bring our AI to a billion Indians,” Sarvam co-founder Pratyush Kumar said during an event in Delhi.
Servum agents are also known as AI models that can complete tasks like large-scale coding or meeting planning autonomously and with minimal human intervention. The company says its agents can boost enterprise automation in one of the world’s fastest-growing economies.
The India-specific models trained from scratch by the startup have been unveiled at a time when the AI race between the US and China is intensifying. The government is funding AI accelerators and incentivizing model makers to launch services so that a country with one of the largest reservoirs of tech talent globally is not left behind.
Sarvam faces tough challenges in fending off global competitors. The startup has received more than $50 million in funding, including from Lightspeed Ventures LLC and Khosla Ventures, and was last valued at about $200 million. That’s tiny compared to Silicon Valley leaders like OpenAI and Anthropic PBC, whose previous valuations were $500 billion and $380 billion, respectively. It is also smaller than companies like French technology leader Mistral AI, which is valued at $13.25 billion and is expanding with local languages in India.
Sarvam promotes its India-first approach and the security it provides by running its AI models from inside the country. The startup’s models are trained on trillions of Indian data sets, exclusively in Indian languages, making it suitable for real-time deployment in the world’s most populous country, not only in local languages but also in mixed languages like Hinglish.
In recent benchmark tests, the startup said its models performed with better accuracy on tasks like optical character recognition for Indian scripts. The Servum Vision model achieved an accuracy of over 84% on document intelligence tasks, outperforming global models hundreds of times larger in size.
“Sovereignty in AI matters much more than building the biggest models,” Vivek Raghavan, Sarvam’s other co-founder, said at the same event.
India is hosting dozens of leading global CEOs, AI founders, country leaders, researchers and policy experts this week as it seeks to position itself as an alternative to the US and China by taking a cost-efficient, language-diverse route to “democratizing AI”.
India’s digital infrastructure has been dependent on foreign technologies for decades, so the launch of the Sarvam model is seen as a step towards developing a “sovereign AI” ecosystem within the country.
Indian AI startups like Sarvam and BharatGen, which also released a series of India-made models this week, are looking to export their AI systems to other developing economies of the world, where neither Chinese nor American models are preferred.
