Bengaluru Startup Sarvam Chosen to Develop India’s First Foundational LLM

In a significant move towards technological self-reliance, the Indian government has selected Bengaluru-based start-up Sarvam to develop the nation’s first homegrown large language model (LLM). This decision comes at a time when global AI landscapes are being reshaped by developments like China’s DeepSeek model. 

Government Support and Strategic Plans

Sarvam was chosen from among 67 applicants and will be supported by the government with vital computing resources to build the AI model from scratch. It is the first start-up to receive incentives under the IndiaAI Mission, a ₹10,370 crore initiative aimed at advancing the country’s AI capabilities.

As part of this support, Sarvam will gain access to 4,000 graphics processing units (GPUs) for a six-month period, provided through companies selected to establish AI data centres in India. A senior official noted that the model would not be open-sourced but would be fine-tuned, particularly for Indian languages.

Speaking on the innovation, IT Minister Ashwini Vaishnaw stated, “This (Sarvam’s) model will have 70 billion parameters and many innovations in programming as well as engineering. With these innovations, a 70 billion parameter (model) can compete with some of the best in the world.”

Development of Model Variants

Sarvam’s project involves the creation of three model variants: Sarvam-Large for advanced reasoning and generation, Sarvam-Small for real-time interactive applications, and Sarvam-Edge for compact on-device tasks. Co-founder Pratyush Kumar highlighted that the model would be capable of reasoning, optimised for voice, fluent in Indian languages, and ready for population-scale deployment.

The company emphasised that the model would be built, deployed, and optimised entirely within India, using local infrastructure and developed by Indian talent. In a press statement, Sarvam stated that the initiative would “promote strategic autonomy, accelerate domestic innovation, and secure India’s leadership in AI for the long term.”

Context and Broader Developments

This advancement occurs as DeepSeek, an open-source, low-cost AI model from China, gains momentum in the global AI arena, influencing companies such as Nvidia due to its cost-efficient yet accurate performance.

Co-founder Vivek Raghavan remarked, “This is a crucial step toward building critical national AI infrastructure. Our goal is to build multi-modal, multi-scale foundation models from scratch. When we do, a universe of applications unfolds. For citizens, this means interacting with AI that feels familiar, not foreign. For enterprises, this means unlocking intelligence without sending their data beyond borders.”

Earlier this year, the government selected 10 companies to supply 18,693 GPUs for developing AI models, exceeding the initial target of 10,000 GPUs under the IndiaAI Mission. The companies include Jio Platforms, Yotta, and Tata Communications, among others.

Read More: India to Unveil Its Own AI Model Within 10 Months, Says Union Minister

Conclusion

Sarvam’s selection for developing India’s first indigenous LLM marks a significant milestone in the country’s pursuit of AI independence. With government backing and an ambitious vision, Sarvam aims to position India as a formidable player in the global AI landscape.

Disclaimer: This blog has been written exclusively for educational purposes. The securities mentioned are only examples and not recommendations. This does not constitute a personal recommendation/investment advice. It does not aim to influence any individual or entity to make investment decisions. Recipients should conduct their own research and assessments to form an independent opinion about investment decisions.