The Impressive TimesNew Delhi, May 30: India’s national AI compute capacity has crossed 34,000 GPUs, a major step toward creating a self-reliant digital ecosystem. Alongside this, three additional startups have been selected to build India’s own Foundation Models under the IndiaAI Mission, further strengthening the country’s artificial intelligence infrastructure.
The announcements were made by Ashwini Vaishnaw, Minister for Electronics & Information Technology, during the ‘IndiaAI – Make AI in India, Make AI Work for India’ event held in New Delhi.
Addressing the event, Vaishnaw urged the newly selected AI teams to strive for top-five global positions in their domains. Reaffirming Prime Minister Narendra Modi’s vision of democratizing technology, he said, “Technology should not be limited to a few. It must be accessible to a larger section of society to create new solutions and unlock opportunities. That is the principle behind the IndiaAI Mission.”
Compute Capacity Expansion
Ashwini Vaishnaw announced the expansion of the national AI compute platform, with an additional 15,916 GPUs added to the existing 18,417 GPUs, bringing the total to 34,333 GPUs. This expansion, supported by leading industry partners including Cyfuture India, Ishan Infotech, Locuz Enterprise Solutions, Netmagic IT Services, Sify Digital Services, Vensysco Technologies, and Yotta Data Services, will power India’s foundational model development and AI research.
Vaishnaw stressed the importance of shared compute infrastructure in enabling equitable access to AI development, calling it a key pillar of India’s AI democratization strategy.
Launch of Three New Indigenous Foundation Models
Under the IndiaAI Foundation Model pillar, 506 proposals were received by April 30, 2025. Based on a multi-stage expert evaluation, the following three have been selected:
Soket AI will build India’s first open-source 120 billion parameter foundation model optimized for Indian languages and sectors like defence, education, and healthcare.
Gnani AI will develop a 14 billion parameter Voice AI foundation model for real-time multilingual speech processing with deep reasoning capabilities.
Gan AI will create a 70 billion parameter multilingual foundation model focused on advanced Text-to-Speech (TTS) systems aiming to surpass current global benchmarks.
These join Sarvam AI, which is already developing a sovereign open-source 120B LLM as part of India’s AI innovation roadmap. Sarvam AI had earlier launched the Sarvam-1 (2B) and Sarvam-M (24B) models.
AI Kosh and Reverse Brain Drain
Vaishnaw shared that 367 datasets have been uploaded to AI Kosh, India’s national AI data repository. He emphasized that the IndiaAI Mission is not only boosting foundational models and compute infrastructure but is also catalyzing reverse brain drain, attracting top talent back to India, and promoting ethical and inclusive AI development.
Cybersecurity Hackathon Winners Announced
Ashwini Vaishnaw also announced the winners of the IndiaAI I4C CyberGuard AI Hackathon, conducted in collaboration with the Indian Cyber Crime Coordination Centre (I4C). The winning teams developed AI-based tools for automated cybercrime complaint classification, trend analysis, and crime pattern recognition using inputs like handwritten FIRs, screenshots, and voice calls.
About IndiaAI
IndiaAI, an independent business division under MeitY, is the implementing body of the IndiaAI Mission. Its mandate includes democratizing AI access, fostering indigenous AI solutions, ensuring responsible use, and positioning India as a global AI powerhouse.
No Comments: