Ola founder Bhavish Aggarwal's AI start-up Krutrim becomes India's first AI unicorn

In a blog post, the company announced that it is the first Indian AI startup to reach a billion-dollar valuation, achieved just one month after introducing a substantial language model.

The  Sanskrit term "Krutrim" translates to "artificial" in English, and the company is not only working on large language models but is also actively involved in the development of data centers.

Krutrim's long-term goal is to produce servers and supercomputers for the AI ecosystem.

In response to the launch of OpenAI's ChatGPT over a year ago, several Indian startups and academic groups are in a race to create large language models in Indian languages, known as Indic LLMs.

This push is part of a global trend where countries aspire to develop their own competitive AI systems instead of relying on technology from the United States or China.

In Europe, investors are heavily investing in Mistral AI in France, which has been valued at $2 billion since its establishment last year. Meanwhile, the United Arab Emirates is showcasing its Falcon model, supported by a government research institute in Abu Dhabi.

India, with its population of 1.4 billion, is concentrating on building smaller, more cost-efficient AI systems. Sarvam, a generative AI startup, recently launched OpenHathi, its first open-source Hindi language model.

The company utilized existing open-source models to build its system.

This announcement closely followed Sarvam raising $41 million in investments from Lightspeed Venture Partners, billionaire Vinod Khosla, and other contributors.

“India has to build its own AI,” Aggarwal, the founder of Indian ride-hailing startup Ola, said in the statement. “We are fully committed towards building the country’s first complete AI computing stack.”

According to the blogpost, “Krutrim is a family of Large Language Models, including Krutrim base and Krutrim Pro which will have multimodal, larger knowledge capabilities, and many other technical advancements for inference. Trained on over 2 trillion tokens, Krutrim accomplishes better performance on multiple well-known, global, LLM evaluation benchmarks including MMLU, HellaSwag, BBH, PIQA and ARC.”

Krutrim model can fluently switch between languages and discuss nuanced topics ranging from poetry in Bengali, to Bollywood movies, to creative masala dosa recipes! read the blog post.

The Krutrim AI model will be available in beta version for consumers in February 2024, the company said.

Additionally, it will also be available as an API for enterprises and developers, seeking to create AI applications.

The company said its “superior linguistic skills make it a valuable tool for a wide range of purposes from education to business communications.”

The company is also working on AI infrastructure to develop indigenous data centers and eventually, server-computing, edge-computing and super-computers.