India is making significant strides in bolstering its artificial intelligence capabilities, with the government broadening its efforts to develop indigenous AI foundation models. Following the earlier selection of Sarvam AI, three additional pioneering teams—Soket AI, Gan AI, and Gnani AI—have now been chosen to spearhead the creation of large-scale AI models. This strategic expansion was announced by IT and Electronics Minister Ashwini Vaishnaw, who underscored the nation's commitment to building a robust AI ecosystem.

These selections are a core component of the ambitious IndiaAI Mission, specifically under its Foundation Model pillar, which has already attracted 506 proposals since its inception. The overarching goal is to establish sovereign AI capabilities by developing foundational models meticulously trained on India-specific data. These models are envisioned to play a crucial role in enhancing governance, improving public service delivery, and fostering greater digital inclusion across the nation, catering to India's unique linguistic and sectoral requirements.

Soket AI has been tasked with developing an open-source foundational model with an impressive 120 billion parameters. This model will primarily focus on addressing India's vast linguistic diversity and is expected to find critical applications in strategic sectors such as defense, healthcare, and education, thereby contributing to national development and security.

Gnani.ai will concentrate on creating a 14 billion parameter Voice AI foundation model. This model aims to provide real-time, multilingual speech processing capabilities coupled with advanced reasoning. Ganesh Gopalan, Co-Founder and CEO of Gnani.ai, expressed that this selection reinforces the company's mission to make technology more inclusive and accessible. "We’re proud to lead the way in developing voice-to-voice large language models for India and the world, because we believe transformative AI must speak the language of the people it serves," he stated.

The third selected entity, Gan.ai, will undertake the development of a 70 billion parameter multilingual model. A key focus for Gan.ai will be on pioneering next-generation text-to-speech capabilities, with the ambitious target of not just meeting but exceeding current global standards in voice synthesis and interaction.

This new cohort of AI developers joins Sarvam AI, which was selected on April 26 to build India’s Sovereign Large Language Model (LLM) ecosystem. Sarvam AI is also working on an open-source 120 billion parameter model designed to support flagship government use cases, including initiatives like “2047: Citizen Connect” and “AI4Pragati,” which aim to enhance citizen engagement and promote progress through AI.

To power these ambitious AI development efforts, the Indian government is significantly scaling up its computational infrastructure. Minister Ashwini Vaishnaw announced a substantial expansion of AI compute capacity, with an additional 15,916 Graphics Processing Units (GPUs) being made available. This augmentation brings the total number of GPUs dedicated to AI research and development to 34,333 (up from the existing 18,417). Vaishnaw emphasized that this enhanced compute facility is crucial for enabling India to develop its AI ecosystem comprehensively. He also highlighted the progress of the AI Kosh, a data repository now holding 367 datasets, and noted the burgeoning app ecosystem around AI. The minister reiterated the focus on the "democratisation of technology" and set a high bar for the selected teams, stating, "Whichever sector they focus on, they must be among the top five in the world."

The expanded GPU resources, totaling 34,333 units, will be accessible to startups and researchers via a common cloud-based platform designed for both model training and inference. This initiative is strongly backed by industry partnerships. To facilitate the supply of this critical GPU infrastructure, seven companies have already submitted commercial offers across various AI compute categories. These companies are Cyfuture India Pvt. Ltd., Ishan Infotech Ltd., Locuz Enterprise Solutions Ltd., Netmagic IT Services Pvt. Ltd., Sify Digital Services Ltd., Vensysco Technologies Ltd., and Yotta Data Services Pvt. Ltd., signaling robust private sector involvement in realizing India's AI aspirations.