The Indian government has unveiled a major initiative to develop homegrown foundational artificial intelligence (AI) models, signalling a significant push in the country’s AI ambitions.
The Ministry of Electronics and Information Technology, through its IndiaAI Mission, has called for proposals to create AI models that are tailored to Indian needs while meeting global standards.
The initiative invites startups, researchers, and entrepreneurs to collaborate on building large multimodal models (LMMs), large language models (LLMs), or small language models (SLMs) trained on Indian datasets. It aims to position India as a key player in the global AI landscape while addressing unique domestic challenges and opportunities.
This proposal comes after Union IT minister Ashwini Vaishnaw said India is set to develop its own generative AI model, aiming to rival global platforms like ChatGPT and DeepSeek. “Very soon, we will have our own LLMs,” Vaishnaw added. The initiative will be driven by the IndiaAI Compute Facility, which has acquired 18,000 GPUs to support the creation of an LLM tailored to the country’s needs.
Notably, the DeepSeek-V3 model was trained on a cluster of 2,048 NVIDIA H800 GPUs with a budget of only $5.576 million.
Vaishnaw recently praised the progress made by DeepSeek, stating that the country will host the Chinese AI lab’s LLM on domestic servers, marking a rare instance of Chinese technology entering India.
“You have seen what DeepSeek has done – $5.5 million and a very powerful model,” he said, addressing the criticism faced by Delhi regarding its comparatively smaller investment in AI, especially in contrast to other nations.
“The DeepSeek models are a step change and will enable faster progress towards even more powerful intelligent systems,” Sarvam AI founder Pratyush Kumar said. He further said that India should “absolutely build its own sovereign models from scratch”.
He emphasised the importance of creating models that “do not compromise between state-of-the-art deep reasoning skills and high proficiency in all major Indian languages”.
The proposed AI models should meet a few key requirements. The models are expected to be trained on diverse, representative Indian datasets, ensuring cultural, linguistic, and contextual relevance. Moreover, companies developing these models must operate under Indian law. Finally, the models should be created to benefit the public while also serving strategic purposes.
Proposals will be evaluated based on criteria such as the team’s capability, approach, data strategy, and feasibility. Proposals will be funded through direct grants or equity-based funding, with the potential for co-financing by industry stakeholders or investors.
Interested applicants must submit proposals via email to tenders@indiaai.gov.in. The call for proposals will remain open until sufficient applications are selected or six months after the announcement. Evaluations will occur monthly.
