SoftBank-backed billionaire to invest $230M in Indian AI startup Krutrim
OLA founder Bhavish Aggarwal is investing $ 230 million in an AI startup that he has founded while the country insisting in a field that is dominated by the US and Chinese companies.
Aggarwal largely finances the investment in Krtrim through his family office, a source that is familiar with the WAN told. In A message on X On Tuesday, Aggarwal said that Krutim wants to attract an investment of $ 1.15 billion by next year. He will try to increase the rest of the capital of external investors, the source said.
The announcement of the financing coincides with a Unicorn startup Krtrim who makes its AI models open source and reveal plans to build what it claims will be the largest supercomputer in India in collaboration with Nvidia.
The laboratory Released Krtrim-2, a parametermodel of 12 billion parameters that has shown strong performance in processing Indian languages. In sentiment analysis Test Krtrim Shared TuesdayIt scored 0.95 compared to 0.70 for competing models, while it achieved a success rate of 80% when generating codes.
The lab has opened various specialized models open, including systems for processing images, speech translation and text search, all optimized for Indian languages.
“We are nowhere close to global benchmarks, but have made good progress in one year,” wrote Aggarwal, whose other companies are supported by Softbank, on X. “By buying our models open, we hope that the entire Indian AI community Works together to create an Indian AI ecosystem of World Class.
The initiative comes if India wants to settle in an artificial intelligence landscape that is dominated by the American and Chinese companies. The recent release of the R1 -“reasoning” model of Deepseek, built on a so -called modest budget, has sent shock waves by the technical industry.
India praised the progress of Deepseek last week and said that the country will organize the great language models of the Chinese AI Lab on domestic servers. Krutim’s cloud arm Deepseek began to offer on Indian servers Last week.
Krutim has also developed its own evaluation framework, BharatbenchTo assess the skill of AI models in Indian languages, with a gap in existing benchmarks that mainly focus on English and Chinese.
The technical approach to the lab includes the use of a context window of 128,000 branches, allowing its systems to process longer texts and more complex conversations. Performance metrics published by the startup showed that Krtrim-2 achieved high scores in grammar correction (0.98) and multi-turn conversations (0.91).
The investment follows the launch of January of Krutrim-1, a parameter system of 7 billion that served as the first major language model of India. The supercomputer implementation with NVIDIA is planned to go live in March, with expansion planned all year round.
