Ola Founder Bhavish Aggarwal To Invest Rs 2,000 Crore In Krutrim, Launches AI Lab

Bhavish Aggarwal, the founder and CEO of Ola, has made a significant announcement regarding his artificial intelligence (AI) startup, Krutrim. He revealed plans to invest Rs 2,000 crore into the venture, with an additional commitment of Rs 10,000 crore expected by next year.

In a recent post on X, Aggarwal celebrated the official launch of the Krutrim AI lab. He noted that although the team has been working on AI projects for over a year, they are now ready to share their research and developments with the open-source community. Aggarwal confirmed the publication of numerous technical reports to foster collaboration.

Krutrim is dedicated to enhancing AI capabilities specifically for India, focusing on improving processes and results related to Indian languages, addressing data scarcity, and taking into account cultural contexts. This initiative aims to create a local AI ecosystem that can compete globally.

This investment and development effort comes in the wake of competition from firms like DeepSeek, a Chinese startup that introduced the DeepSeek-R1 model, claiming it can match the performance of leading Western AI technologies at a lower cost.

Technology Partnerships and Future Plans

Aggarwal also announced a partnership with Nvidia to deploy the GB200 GPU card, part of the Nvidia Blackwell platform. The expectation is to have this new system operational by March, with aspirations for it to become the largest supercomputer in India by the end of the year.

While recognizing that India has not yet reached global benchmarks in AI, Aggarwal expressed optimism about the progress made in just one year. He believes that by open-sourcing their models, Krutrim can encourage the Indian AI community to come together and develop a world-class AI ecosystem. In his post, he remarked on the importance of learning and growing at a steady pace.

New AI Models and Benchmark Developments

Alongside the AI lab launch, Aggarwal introduced an upgraded version of Krutrim's large language model, named Krutrim 2, following the earlier release of Krutrim 1 in January 2024. The new model was presented along with several additional models developed from Krutrim 1.

Among these is Chitrarth 1, which serves as a vision language model adept in interpreting images and documents; Dhwani 1, designed for speech-related tasks such as translations; Vyakhyarth 1, an Indic embedding model useful in applications like search and retrieval-augmented generation; and Krutrim Translate 1, focused on text-to-text translations.

Additionally, the newly developed “BharatBench,” a global benchmark tailored for Indic performance, was revealed, alongside a collection of technical reports and papers contributing to the field.

Furthermore, Krutrim aims to provide large language model (LLM) services specifically for the Kumbh Sah’AI’yak App, showcasing their commitment to integrating AI within the broader cultural frameworks of India.

Ola, AI, Investment