Krutrim pivots to cloud after layoffs, as AI ambitions face headwinds

Krutrim’s move toward cloud services, after multiple rounds of layoffs and few major product updates, has turned into a blunt lesson in the economics of India’s AI race. Bhavish Aggarwal launched the company in 2023, and by January 2024 it had become India’s first AI unicorn after raising about $50 million at a $1 billion valuation.

The early pitch was ambitious. In May 2024, Aggarwal said Ola would move its workloads off Microsoft Azure and onto Krutrim Cloud within a week, a declaration that framed the startup as a potential domestic alternative to foreign infrastructure. By August, he was talking even bigger, laying out plans for a 1-gigawatt data center by 2028, up from 20 megawatts at the time. In February 2025, he announced a fresh Rs 2,000-crore investment in Krutrim and said Ola would commit Rs 10,000 crore by the following year, while unveiling Krutrim 2, an AI lab and India’s first GB200 deployment.

The market, however, proved less forgiving than the pitch. By June 2025, reports said Krutrim’s large language models and cloud products were getting a tepid response from some founders and developers, with friction ranging from documentation problems to persistent sign-in errors. Kruti, the consumer AI assistant Krutrim launched in June 2025 inside the Ola app, did not break out as a mass product either. Sensor Tower figures cited in reports put downloads in the low hundreds of thousands, ranging from about 207,000 to 270,000 since launch.

The staffing changes that followed underscored how hard it is to turn model-building into a durable business. Krutrim laid off more than 100 employees in July 2025, mostly from its linguistics division, and by September reports said a third round of cuts had pushed total departures past 200. That came as the company’s broader roadmap, including multibillion-token Indic-language models and work on Krutrim 3, was said to be slowing or stalled.

By April 2026, reports said Kruti had been quietly shut down less than a year after launch. For India’s AI sector, the sequence is a sharp reminder that national ambition is not enough on its own. Frontier-model development demands deep capital, scarce talent, heavy infrastructure and products that can win users, and Krutrim’s pivot suggests cloud services may be the more realistic path for a startup still trying to prove it can build at scale.