India's GenAI Darling Changes Course
Krutrim, the startup that became India's first generative AI unicorn, is quietly stepping back from its original ambition of building large-scale AI models to compete with OpenAI and Google — and pivoting instead toward cloud services.
Founded by Ola co-founder Bhavish Aggarwal, Krutrim burst onto the scene with bold promises: a homegrown Indian AI model trained on local languages and cultural context, purpose-built for the world's most populous country. It reached unicorn status faster than almost any AI startup in the region. But the latest moves tell a more sobering story.
Layoffs and Limited Progress
The company has gone through a round of layoffs and has offered limited product updates over the past several months — both signs that the original model-building roadmap has hit serious friction. According to a TechCrunch report, Krutrim is now repositioning itself primarily as a cloud infrastructure provider rather than a frontier AI lab.
That's a significant shift. Cloud services are a competitive but more straightforward business — offering compute, storage, and hosting to enterprises. Building and maintaining a cutting-edge AI model, on the other hand, requires continuous massive investment in data, compute, and world-class research talent. It's a race that even well-funded Western startups are struggling to win.
The Broader Challenge for AI Startups Outside the US
Krutrim's pivot reflects a tension playing out globally: the gap between AI ambition and AI economics. Training large language models requires enormous amounts of GPU compute — hardware that is expensive, in short supply, and largely controlled by a handful of American companies.
For startups in emerging markets like India, this creates a compounding challenge. The cost of frontier model training is already prohibitive; without the same access to capital markets, hyperscaler partnerships, or government subsidies that US labs enjoy, sustaining a genuine AI research operation is extraordinarily difficult.
Several Chinese AI labs have faced similar headwinds, though China's government has thrown significant state support behind its domestic AI industry. India's approach has been less interventionist, leaving companies like Krutrim to navigate the economics largely on their own.
Cloud as a Pragmatic Path Forward
Pivoting to cloud services isn't a retreat so much as a recalibration. India's cloud infrastructure market is growing rapidly, and there is genuine demand for domestic providers who can offer data residency, regional compliance, and pricing in local currency. If Krutrim can carve out a niche in sovereign cloud or AI-focused infrastructure for Indian enterprises, there's a real business there.
But it's a very different business than being the company that builds the AI models powering India's digital future — which was the original pitch.
What Comes Next
For now, Krutrim remains a company to watch. It still carries the symbolic weight of being India's first GenAI unicorn, and Aggarwal is a high-profile founder with strong industry connections. Whether the cloud pivot stabilizes the business or becomes a longer-term repositioning remains to be seen.
What's clear is that the global AI landscape is separating into two tiers: a small group of labs with the capital to race for frontier models, and a much larger group finding more sustainable paths in application, tooling, and infrastructure.
Krutrim, it seems, is joining the latter.
Source: TechCrunch
