Krutrim, India's First AI Unicorn, Bets on Cloud Services

Krutrim, the first Indian AI startup to reach unicorn status (a valuation above $1 billion), has announced a radical change of strategy. Instead of developing its own generative language models, the company is refocusing on cloud services and applications. The shift followed a series of layoffs, limited product updates and a reassessment of market economics.

Why models are no longer the priority

Building a competitive foundation AI model costs enormous sums. It requires:

• Expensive chips and compute infrastructure
• Huge, high-quality datasets
• Recruiting top ML research talent

OpenAI, Google and Meta* have already invested billions of dollars and control the main market for foundation models. For a startup, even one that has reached unicorn status in India, competing in this field proved impractical. Local investment is limited, resources are expensive, and the window of opportunity narrows every month.

Krutrim tried to build a model focused on Indian languages and the context of local users. But the constant need for capital spending, a shortage of world-class specialists and the need to compete with entrenched giants made it hard to scale. As a result, the company cut staff and reassessed its priorities.

Cloud services as a more adaptable path

Krutrim's new focus is building cloud platforms and applications on top of existing models, including open models and APIs from major providers. This approach has clear advantages:

• Less capital investment in infrastructure
• Easier to scale with demand
• Faster adaptation to regional needs
• The team concentrates on UX, integration and vertical solutions

India's market badly needs affordable AI tools for small and medium businesses, but global platforms overlook local specifics: languages, regulation, payment systems and cultural nuances.

A lesson for India's AI industry

Krutrim's story reflects a broader reality: in an era of big models and big spending, even successful startups must find a niche between ambition and economics. The ones doing better use existing foundation models and build solutions for specific verticals and markets on top of them.

The real value in AI lies not in building your own base model, but in

properly integrating and adapting existing technologies to users' specific problems.

This is a signal to investors: there is no longer a need to fund everyone who wants to build their own model. Success goes to those who find the right application.

What this means

Krutrim's move is not a defeat but a rethink of its business model. For India and other emerging markets, it may open new opportunities in the AI sector by shifting the focus from ambitious fundamental research to practical, scalable applications. Companies that learn to integrate AI well into solving real problems, in finance, healthcare, logistics and e-commerce, will create more value than those trying to compete with the giants in base-model development.

*Meta has been designated an extremist organization and is banned in Russia.