Krutrim, Ola’s AI arm, has launched sovereign AI infrastructure, built a multilingual LLM with 2 trillion tokens, and unveiled Kruti, India’s first agentic AI assistant. With $230M backing and public sector partnerships, Krutrim positions itself as India’s answer to global AI leaders, despite cultural and execution challenges.
India’s AI market is projected to reach USD 8 billion by 2025, with annual growth rates above 35%. Government-backed initiatives such as AI skilling, data governance policies, and semiconductor manufacturing have laid the foundation for a domestic ecosystem. Against this backdrop, Krutrim—Ola’s AI venture—aims to become India’s most comprehensive AI player, spanning infrastructure, language-first large language models, and consumer-facing products.
What differentiates Krutrim is its India-first approach. While global AI platforms dominate in English and Western contexts, Krutrim focuses on India’s 22 official languages, cost-sensitive infrastructure, and everyday use cases like digital payments, mobility, and citizen services.
Krutrim is building a sovereign AI cloud, designed to handle AI training and inference workloads without dependence on foreign hyperscalers. The stack includes:
This sovereign approach aims to address concerns around data privacy, cost of compute, and technological dependence—crucial issues for a country of 1.4 billion people.
In 2025, Krutrim launched Kruti, an AI assistant that goes beyond chat. Unlike traditional chatbots, Kruti is agentic—it can plan, reason, and complete tasks across apps. Examples include:
Kruti represents the next stage of AI adoption in India—moving from Q&A-style bots to full-service digital agents. Its integration with Ola’s ecosystem (mobility, food, payments) gives it a natural head start.
At the core of Kruti lies Krutrim LLM, trained on over 2 trillion tokens across India’s major languages. This model outperforms leading global models in Indic languages and introduces retrieval features for real-time updates.
Key highlights:
This multilingual capability positions Krutrim as the go-to platform for vernacular internet users, who make up a majority of India’s digital population.
Krutrim became India’s first AI unicorn in 2024 with a $50 million raise at a $1 billion valuation. By early 2025, founder Bhavish Aggarwal committed $230 million from his family office, with plans to raise $1.15 billion by 2026 to fuel data centers and chip design.
The company also acquired BharatSahayak, an AI-driven citizen services platform, to embed its AI into public-facing utilities such as governance, education, and healthcare. This positions Krutrim as not just a consumer tech company but also a national infrastructure player.
Despite its promise, Krutrim has faced challenges:
These issues underscore the human cost of hyper-growth in AI startups and highlight the need for better governance and employee support systems.
Focus Area | Krutrim’s Strengths | Key Risks & Challenges |
---|---|---|
Infrastructure | India-first sovereign cloud, chip R&D, data centers | High capital intensity, competing with hyperscalers |
AI Models | 2T-token LLM, multilingual performance edge | Needs to match global frontier models |
Assistant (Kruti) | Agentic, task completion, multilingual support | Adoption curve, user trust, technical maturity |
Funding | Backed by founder capital + unicorn valuation | Execution risk, dependence on future fundraising |
Governance | Public-sector focus via BharatSahayak | Regulatory complexity, cultural controversies |
India’s AI race is intensifying. Startups like Sarvam AI and global entrants like OpenAI, Google, and Anthropic are competing for enterprise and consumer markets. India’s differentiator is its language diversity, cost-sensitive markets, and government support for sovereign AI infrastructure.
Krutrim’s model—combining AI infrastructure, consumer apps, and national service integration—is unique. If executed well, it can make India less dependent on foreign AI providers and potentially export Indic AI solutions to other emerging markets.
Krutrim is one of the boldest bets in India’s AI story—combining sovereign cloud infrastructure, a multilingual LLM, and an agentic assistant in one package. It represents a vision of AI built for India’s 1.4 billion people, tailored to their languages and everyday needs.
But ambition comes with challenges. To succeed, Krutrim must balance scaling infrastructure with cultural sustainability, capital efficiency, and user trust. If it can achieve this, it may not only define India’s AI trajectory but also emerge as a model for AI development in other diverse, emerging markets.