Unlike ChatGPT and Claude, Sarvam AI has introduced an artificial intelligence approach that is better suited for India. At the India AI Impact Summit 2026 in New Delhi, which served as a showcase for Prime Minister Narendra Modi’s initiatives to position his nation as a leader in the developing technology, the Bengaluru-based AI startup unveiled two models. The company claims that Sarvam’s models, which are designed to be operated by voice commands and are available in 22 Indian languages, will provide it a competitive edge in the 1.45 billion-person nation where the great majority of people are illiterate.
“Today we show we can bring our own AI to a billion Indians,” said Sarvam co-founder Pratyush Kumar during an event in Delhi.
Additionally, Sarvam provides what are known as agentic AI models, which are capable of performing tasks like meeting planning and coding mostly on their own with little assistance from humans. According to the company, enterprise automation in one of the fastest-growing economies in the world might be fueled by its agents. As the US-China AI race heats up, the startup has shown India-specific models that were created from scratch. To ensure that the nation, which is home to one of the world’s largest pools of technical talent, doesn’t fall behind, the government is supporting AI accelerators and encouraging model makers to start offering their services.
Sarvam has a difficult time keeping up with international rivals. The business, which was recently valued at over $200 million, has raised more than $50 million in funding, notably from Lightspeed Ventures LLC and Khosla Ventures. Silicon Valley titans like OpenAI and Anthropic PBC, which were recently valued at $500 billion and $380 billion, respectively, dwarf that. It’s also little in comparison to businesses like Mistral AI, a $13.25 billion French technology pioneer that is growing in India using local languages.
Sarvam touts its India-first approach and the security it offers by running its AI models from inside the country. The startup’s models are trained on trillions of Indian data sets, particularly those in Indian languages, making it suitable for real-time deployment at scale in the world’s most populous country not just in local languages but also mixed languages such as Hinglish.
In recent benchmark tests, the startup said its models performed with superior accuracy on tasks like optical character recognition for Indian scripts. The Sarvam vision model achieved an accuracy of over 84% on document intelligence tasks, eclipsing global models hundreds of times larger in size.
“Sovereignty matters much more in AI than building the biggest models,” said Sarvam’s other co-founder, Vivek Raghavan, at the same event
