Ola Founder's AI Startup Krutrim Launches DeepSeek R1 Model at Groundbreaking Price
Krutrim, the artificial intelligence startup co-founded by Ola's Bhavish Aggarwal, has announced the deployment of the DeepSeek R1 671B AI model in India at a disruptive price point of Rs 1 per million tokens for February, marking a significant move to democratize AI access in the country.
The company has achieved a global first by deploying DeepSeek-R1 671B on Nvidia's H100 graphics processing units through Krutrim AI Cloud. This pricing strategy represents a dramatic reduction compared to competitors, with similar models like Llama-3-8B-Instruct costing Rs 16.60 per million tokens, Google's Gemma-27B at Rs 66.40, and Hugging Face's M4/idefics2-8B at Rs 16.60.
Aggarwal emphasized the need to accelerate India's AI journey to a new level; so they partnered with DeepSeek. The initiative aligns with Krutrim's broader mission to foster AI innovation and adoption aamongIndian businesses and developers.
In a significant infrastructure push, Krutrim has announced plans to deploy Nvidia's GB200 GPUs by March 2025, aiming to create what Aggarwal describes as "the largest supercomputer in India by the end of the year." The company has also launched the Krutrim AI Lab with an initial investment of ₹2,000 crore, expected to increase to Rs 10,000 crore by next year.
Krutrim, whose Sanskrit name means "artificial," launched in April 2023 and achieved unicorn status by January 2024. The company was co-founded by Aggarwal and Krishnamurthy Venugopala Tenneti, a board member of ANI Technologies, which owns Ola and Ola Electric.
The startup's commitment to open-sourcing much of its work in 2024 reflects its vision to build a collaborative AI ecosystem in India. Through local hosting of robust models and focused development on Indian languages and cultural context, Krutrim aims to address the unique challenges of data scarcity and linguistic diversity in the Indian market.
India's AI development landscape is all set to flourish with Krutrim, which offers affordable access to advanced AI models with substantial infrastructure investment.