SoftBank-backed billionaire to invest $230M in Indian AI startup Krutrim | TechCrunch
Feb 04, 2025 - techcrunch.com
Ola founder Bhavish Aggarwal is investing $230 million into Krutrim, an AI startup he founded, as part of India's efforts to establish itself in the AI sector dominated by U.S. and Chinese firms. The investment is primarily financed through Aggarwal's family office, with plans to raise an additional $1.15 billion from external investors. Krutrim has made its AI models open source and announced a partnership with Nvidia to build India's largest supercomputer. The startup released Krutrim-2, a 12-billion parameter language model optimized for Indian languages, showing strong performance in sentiment analysis and code generation tasks. Krutrim has also developed BharatBench, an evaluation framework for AI models in Indian languages, and uses a 128,000-token context window for handling complex texts.
The initiative aligns with India's ambition to enhance its presence in the AI landscape, further highlighted by the recent release of DeepSeek’s R1 model. Krutrim's cloud arm has begun hosting DeepSeek on Indian servers. The startup's technical advancements include high scores in grammar correction and multi-turn conversations. The investment follows the earlier launch of Krutrim-1, India's first large language model, with the supercomputer deployment set for March. The move aims to foster collaboration within the Indian AI community to build a world-class AI ecosystem.
Key takeaways:
Ola founder Bhavish Aggarwal is investing $230 million into his AI startup Krutrim, aiming to raise a total of $1.15 billion by next year.
Krutrim has open sourced its AI models and plans to build India's largest supercomputer in partnership with Nvidia.
The startup has developed Krutrim-2, a 12-billion parameter language model optimized for Indian languages, achieving high performance in sentiment analysis and code generation tasks.
Krutrim introduced BharatBench, an evaluation framework for AI models' proficiency in Indian languages, addressing a gap in existing benchmarks.
