Sarvam Unveils New AI Models, Hires Former xAI Researcher Devendra Chaplot  

Sarvam AI has announced a series of product updates at its Epoch conference while also strengthening its research team with the appointment of Devendra Chaplot, an AI researcher who most recently worked at Elon Musk’s xAI. 

An IIT Bombay alumnus, Chaplot holds over a decade of experience in AI, robotics and machine learning. He has previously held research roles at Facebook AI Research (FAIR), Mistral AI and Thinking Machines Lab. He has also received the Facebook Fellowship and multiple international AI research awards for his work in embodied AI and autonomous navigation. 

The appointment comes weeks after Sarvam raised USD 234 million in the first close of its Series B funding round to accelerate frontier model research, expand compute infrastructure and strengthen its enterprise and government AI offerings.

The twin developments come as the Bengaluru-based startup expands its efforts to build foundation models and AI infrastructure tailored for India. 

At the two-day Epoch event, Sarvam introduced a new suite of Indian language AI models and developer tools aimed at improving multilingual AI capabilities. The company also showcased updates to its existing models, new enterprise AI products and infrastructure designed for developers and businesses building AI applications. The conference featured technical workshops, product demonstrations and research presentations focused on India’s AI ecosystem.

Speaking at Epoch, Pratyush Kumar, Co-founder of Sarvam AI, said: “The compute cost was definitely lesser than what DeepSeek took by a significant margin. In fact, the whole compute was about USD 20 million, or a little bit less. And we trained three models.”

He added: “It’s not about one run that it costs. It’s a lot of exploratory work. It’s a lot of data creation work that gets layered. But the bottom line is, it is not billions of dollars. It’s a single-digit million dollar model.”