Building India's Own LLMs: The Path to AI Independence

Building India’s Own LLMs: The Path to AI Independence

AI is transforming the world, and India stands at a crossroads—will we lead or follow? With homegrown Large Language Models (LLMs) emerging, we have an unprecedented opportunity to shape AI to reflect our diverse linguistic and cultural heritage. The stakes couldn’t be higher.

History favors the bold. DeepSeek, a Chinese startup, built an advanced AI model for just $6 million, proving that success in AI isn’t about money alone—it’s about vision and execution. For India, this is crucial. With 143 crore people speaking dozens of languages, we need AI that serves our population, not just English speakers.

The government has allocated ₹10,370 crores under the IndiaAI Mission, anticipating 27 lakh AI jobs by 2028 and a $500 billion boost to the economy. But this potential remains unrealized without LLMs built for India.

Startups like Sarvam AI and Krutrim are stepping up, but challenges remain—particularly in data. High-quality Indian language datasets are scarce, with archives and newspapers largely undigitized. Infrastructure, however, is improving, with major cloud investments and 18,693 GPUs approved for AI research.

Our Founder and CEO, Bhavesh Goswami insightfully notes, “Understanding the core of hardware and its interplay with code isn’t just a skill—it’s an edge in today’s competitive landscape.” India must build AI that not only understands its languages but also safeguards its identity.

The question remains: Will we seize this moment, or will we let others define AI’s future for us? Read more to know what Bhavesh says in The Economic Times.