Sarvam, an Indian AI startup focused on building models for local languages and users, officially launched its Indus chat app on Friday for web and mobile platforms. The move positions the company in a fast-growing generative AI market, currently dominated by global players like OpenAI, Anthropic, and Google.
India has emerged as a major hub for AI adoption. OpenAI CEO Sam Altman recently noted that ChatGPT sees over 100 million weekly active users in India, while Anthropic reported that the country accounts for 5.8% of Claude usage, second only to the U.S.
Indus and the Sarvam 105B Model
Indus acts as the chat interface for Sarvam 105B, the company’s newly unveiled 105-billion-parameter large language model. The app’s launch follows closely after Sarvam introduced both the 105B and 30B models at the India AI Impact Summit in New Delhi earlier this week.
At the summit, Sarvam also highlighted enterprise initiatives and hardware plans and announced partnerships with companies including:
HMD: Bringing AI to Nokia feature phones.
Bosch: Developing AI-enabled automotive applications.
Features and Limitations
Currently available in beta on iOS, Android, and the web, Indus allows users to type or speak queries, receiving responses in both text and audio. Users can sign in via phone number, Google account, or Apple ID, though access is presently limited to India.
Some early limitations include:
No option to delete chat history without deleting the account.
Users cannot turn off the reasoning feature, which may slow responses.
Access may be restricted during the gradual expansion of compute capacity.
Sarvam co-founder Pratyush Kumar shared on X:
“We’re gradually rolling out Indus on a limited compute capacity, so you may hit a waitlist at first. We will expand access over time.”
Funding and Growth
Founded in 2023, Sarvam has raised $41 million from investors including Lightspeed Venture Partners, Peak XV Partners, and Khosla Ventures. The startup is focused on creating large language models tailored to the Indian market, aiming to compete with global AI offerings while addressing local language and user needs.