India’s first AI unicorn Krutrim bets on cloud services

Krutrim, India's first AI-startup to achieve unicorn status (valuation exceeding one billion dollars), has announced a drastic shift in strategy. Instead of developing its own generative language models, the company is reorienting towards cloud services and applications. This pivot came after a series of layoffs, limited product updates, and a reassessment of market economic realities.

Why Models Are No Longer a Priority

Creating a competitive foundational AI model requires enormous expenditures. It requires:

• Expensive chips and computing infrastructure
• Huge high-quality datasets
• Attracting top talent in ML research

OpenAI, Google, and Meta have already invested billions of dollars and control the primary market for base models. For a startup — even one that achieved unicorn status in India — competing in this field proved impractical. Local investment opportunities are limited, resource costs are high, and the window of opportunity closes with each passing month.

Krutrim attempted to create a model focused on Indian languages and the context of local users. However, the constant need for capital expenditures, a shortage of world-class specialists, and the necessity to compete with established giants made scaling difficult. As a result, the company implemented layoffs and reassessed its priorities.

Cloud Services as a More Adaptive Path

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

• Lower capital investments in infrastructure
• Easier scaling depending on demand
• Faster adaptation to regional needs
• Team focuses on UX, integration, and vertical solutions

The Indian market has an acute need for accessible AI tools for small and medium businesses, but global platforms don't account for local specifics: languages, regulation, payment systems, cultural nuances.

A Lesson for India's AI Industry

Krutrim's story reflects a broader reality: in the era of large models and large expenses, even successful startups must find a niche between ambitions and economics. Those who succeed are those who use existing foundational models and build solutions on top of them for specific verticals and markets.

True value in AI lies not in creating one's own base model, but in

proper integration and adaptation of already existing technologies to specific user problems.

This is a signal for investors: in AI, it's no longer necessary to finance everyone who wants to create their own model. Success comes to those who find the right application.

What This Means

Krutrim's transition is not a defeat, but a reimagining of the business model. For India and other developing markets, this could open new opportunities in the AI sector, shifting focus from ambitious foundational research to practical, scalable applications. Companies that learn to integrate AI well into solving real problems — in finance, healthcare, logistics, e-commerce — will create greater value than attempts to compete with giants in developing base models.