The founder of Ola Bhavish Aggarwal is investing $ 230 million in Amnesty International’s emerging company, while the country is pressuring to establish itself in a field dominated by American and Chinese companies.
AGGARWAL funds the investment in KRUTRIMI to a large extent through his family’s office, a source familiar with the matter said. in Posted on X Tuesday, AGGARWAL said that Krutrim seeks to attract an investment of $ 1.15 billion by next year. The source said that he would seek to raise the remaining capital of external investors.
The financing announcement coincides with the start of the KRUTRIMID in Unicorn Models Open Models and unveiled plans that it claims will be the largest collection of superolo computers in India in partnership with NVIDIA.
the laboratory KRUTRIM-2, a 12 billion teacher language model that has shown a strong performance in the treatment of Indian languages. In KRUTRIM Tuesday joint0.95 compared to 0.70 for competing models, with a success rate of 80 % in the code generation tasks.
The laboratory has opened many specialized models, including image processing, speech translation and text search systems, all of which are improved for Indian languages.
“We are anywhere close to global standards so far, but we have made good progress in one year. To create an ecosystem of Indian artificial intelligence in the world.
This initiative comes at a time when India seeks to establish itself in an artificial intelligent scene dominated by Chinese companies and companies. The latest version of the “Thinking” model from Deepseek, which was built with a modest budget, sent shock waves through the technology industry.
India praised last week the progress of Depsic and said that the country will host large languages for the Chinese artificial intelligence laboratory on local servers. Crotreim cloud arm I started to present Deepseek on Indian servers Last week.
KRUTRIM has also developed its evaluation framework, BharatbenchTo assess the efficiency of artificial intelligence models in Indian languages, and to address a gap in current standards that focus mainly on English and Chinese.
The technical approach to the laboratory includes the use of a window of 128,000, which allows its systems to deal with long texts and the most complex conversations. The performance measures published by the KRUTRIM-2 Playing Start showed high degrees in the correction of the rules (0.98) and multiple conversations (0.91).
The investment follows the launch of the Krutrim-1 in January, a 7-billion parameter system, which was the first major linguistic model in India. The super computer is scheduled to be deployed with NVIDIA in March, with planning to expand throughout the year.
