India’s tech scene just got another player in the generative AI game. Sarvam, a Bengaluru-based startup, has launched Indus, a chat app that’s aiming to take on the likes of ChatGPT, Claude, and Google’s AI offerings. The timing is interesting, especially when you consider that India has quietly become one of the most competitive markets for AI adoption right now.
The app runs on Sarvam’s freshly announced 105B model, a massive language model built specifically with Indian users in mind. We’re talking about an AI system designed to understand local languages and the nuances of the Indian market. That’s the real differentiator here, not just another ChatGPT clone with a new interface.
Why India Matters (and Why Global AI Companies Know It)
The numbers tell a compelling story. Sam Altman recently dropped that ChatGPT has crossed 100 million weekly active users in India alone. That’s not a rounding error or a vanity metric. That’s a massive chunk of the world’s population actively engaging with AI tools. Anthropic’s Claude isn’t far behind either, grabbing 5.8% of its global usage from India, second only to the U.S.
This kind of traction has naturally caught the attention of policymakers and local entrepreneurs alike. India has been increasingly vocal about wanting homegrown technology solutions that keep data and infrastructure within the country’s borders. It’s part of a larger push for digital sovereignty that we’re seeing across different nations.
The Indus App: What’s Actually There
So what does Indus actually do? It’s a chat interface where you can type or speak, and get responses back in both text and audio. Pretty standard stuff on the surface. The sign-in process is straightforward enough with phone number, Google, Microsoft, or Apple ID options. Currently it’s in beta across iOS, Android, and web, though it’s locked to India for now.
But here’s where it gets real: the app comes with some rough edges that feel very much like early-stage software. You can’t delete your chat history without nuking your entire account. That’s not great for privacy-conscious users. The reasoning feature can’t be turned off, which sometimes makes responses slower. And Sarvam’s being transparent about the fact that compute capacity is limited, so you might hit a waitlist.
That last bit matters more than you’d think. Sarvam co-founder Pratyush Kumar was honest about it on X, essentially saying they’re rolling this out slowly and asking for feedback. That’s refreshing compared to the “move fast and break things” mentality that’s defined so much of tech.
The Bigger Picture: Building Indian AI Infrastructure
This isn’t just about launching another chat app. Sarvam announced its 105B and 30B models just two days earlier at the India AI Impact Summit in New Delhi. The company’s also been busy with enterprise partnerships, including work with HMD to bring AI to Nokia feature phones and Bosch for automotive applications. That’s a much broader vision than just capturing chat users.
The startup has raised $41 million so far from serious investors like Lightspeed Venture Partners, Peak XV Partners, and Khosla Ventures. These aren’t small-time players betting on India’s AI scene. They’re institutional money that believes there’s real value in building domestic alternatives to global platforms.
The broader context is that a small but growing cluster of Indian startups are now trying to create homegrown AI infrastructure. It’s part of a global trend where countries are increasingly skeptical about relying solely on American and European tech companies for critical infrastructure. Whether that’s wise policy or protectionism depends on who you ask, but the market incentives are clearly there.
The Real Test Ahead
What happens next will be interesting to watch. Indus needs to move beyond early adopters who’ll tolerate beta limitations. It needs to prove that a locally-built chat app can actually compete on performance, not just on the feel-good factor of supporting Indian innovation. The app’s current limitations suggest that Sarvam still has significant work to do, though that’s pretty typical for a product in its first weeks of public availability.
The question isn’t really whether Indus can beat ChatGPT globally. It probably can’t, at least not in the short term. The question is whether it can become the default choice for Indian users who want to keep their interactions within Indian-controlled infrastructure, and whether that market alone is large enough to sustain the company.
Given that ChatGPT has 100 million weekly active users in India, there’s clearly an enormous addressable market for any player that can execute well. But execution at scale is where most startups stumble.
