Lenovo and Krutrim Collaborate to Create India's Fastest Supercomputer

In addition to developing India's largest supercomputer, Ola's artificial intelligence (AI) initiative, Krutrim, announced on 6 March that it is collaborating with Lenovo to develop Krutrim 3, a 700-billion-parameter large language model (LLM). At the Lenovo Tech World India 2025 in Mumbai, Ola's chief information officer, Navendu Agarwal, stated that the company is extremely proud to be working on a larger model, Krutrim 3, which will be a 700-billion-parameter model and be the company's response from India to demonstrate that the country can build the best. The company is thrilled to be working with Lenovo on this project. The news comes after Bhavish Aggarwal, the creator of Ola, pledged to invest INR 2,000 crore in Krutrim, with ambitions to raise that amount to INR 10,000 crore by the following year.

How it can Benefit the Nation?

Ola has 700 employees committed to developing its full-stack AI solutions, according to Agarwal. In late 2023, the business released its first model, Krutrim 1, which had seven billion parameters. It was followed by Krutrim 2, which had twelve billion parameters, and both were open-sourced. He underlined that one of Krutrim's main areas of interest is to fill India's gap in hyperscale infrastructure and sovereign cloud. "We are quite pleased to announce that, in collaboration with Lenovo, we are constructing India's largest supercomputer, which will be powered by this cloud." He added, "We are creating our own chip and the foundation models."

The development coincides with a growing race among countries to develop effective and affordable sovereign AI models, especially in the wake of DeepSeek, a Chinese AI business. On March 6, Bhavish Aggarwal asked people on the Twitter site X for ideas for a name for the supercomputer and wrote that Lenovo and Kirtrim were building India's biggest supercomputer. "To introduce the AI revolution to India on a large basis, much more is required! I'm all in to see this through to completion," he wrote.

Krutrim Developments Till Now

Aggarwal declared in February that Krutrim would make Krutrim 2 open-source, along with a number of other models, such as Dhwani 1, a voice-language model that can do tasks like speech translation, and Chitrarth 1, a vision-language model based on Krutrim 1. Additionally, he had declared that Krutrim is working with the American chip giant Nvidia to deploy GB200 superchips. In order to improve data privacy and lower the cost of AI training, Krutrim had previously installed sophisticated AI models from DeepSeek on its domestic servers. India is also seeing an increase in the drive for autonomous AI capabilities. Tata Sons chairman Natarajan Chandrasekaran emphasised the significance of building up local AI infrastructure last week at an event in Mumbai, cautioning that without it, India's activities, languages, and cultures run the risk of being processed by AI systems that do not completely comprehend the nation.