Ola founder Bhavish Aggarwal’s startup Krutrim becomes India’s first AI unicorn
Ola founder Bhavish Aggarwal recently revealed that his artificial intelligence startup Krutrim has raised $50 million in funding, leading to his company touching $1 billion valuation. Krutrim, through the latest funding, has become India’s first-ever AI unicorn company.
Aggarwal said that Krutrim, in its latest round of funding, ended up raising $50 billion from investors including Matrix Partners India. Krutrim ended up hitting the billion-dollar mark in valuation just a month after its debut as a large language AI model.
Krutrim, which translates to “artificial” in Sanskrit, is also in the process of developing data centres across the country, aiming to create super computers and supercomputers for the AI ecosystem developing in India, the company said in a blogpost.
The company said in its post that it had plans to make a beta version of its eponymous chatbot available to consumers next month. Krutrim will also roll out APIs to developers and enterprises in the coming months.
It must be noted that Matrix Partners India, which led the latest round of funding for Krutrim, is one one of the prime investors in the other two startups of Bhavish Aggarwal – Ola and Ola Electric.
Aggarwal said in a statement, “We are thrilled to announce the successful closure of our first funding round, which not only validates the potential of Krutrim’s innovative AI solutions but also underscores the confidence investors have in our ability to drive meaningful change out of India for the world.”
What is Krutrim AI?
Krutrim is a large language model (LLM) which has been trained on more than 2 trillion ‘tokens’; these tokens are sub-words used in conversations. The AI platform is also set to have two classifications.
Krutrim was introduced last month while Krutrim Pro is set for release in 2024, with advanced capabilities for problem-solving and task execution. The project team of the company also claims that Krutrim is larger than GPT-4.
This LLM uses a custom tokenizer to interpret the languages and scripts. When compared to other open-source LLMs trained with similar data volumes, it outperformed those on a range of industry-standard benchmarks, the company claims.