Sovereign AI models available at summit, huge response to event: Vaishnaw
At the India AI Impact Summit 2026 (Bharat Mandapam, New Delhi, February 16–21, 2026), IT Minister Ashwini Vaishnaw announced that India's sovereign AI models were available and demonstrated at the summit
Three sovereign AI models were unveiled: Sarvam AI's large language models (30B and 105B parameter models, trained entirely in India), BharatGen's Param2 17B multilingual foundational model, and Gnani.ai's Vachana text-to-speech model (supporting 12 Indian languages)
The government announced plans for significant compute infrastructure expansion: an additional 20,000 GPUs to be added to India's sovereign compute pool (which already exceeds 38,000 GPUs provisioned under the IndiaAI Mission)
A secure GPU cluster of 3,000 next-generation GPUs is being built for sovereign and strategic applications
Reliance and Adani together pledged a combined $210 billion in investment in domestic AI and data infrastructure
IndiaAI Mission — Architecture and Scale
The IndiaAI Mission was approved by the Union Cabinet in March 2024 with an allocation of ₹10,000 crore. It is the Government of India's flagship programme to establish end-to-end sovereign AI capability — from compute infrastructure to foundational model development and application deployment. The mission operates through seven pillars: Compute, Foundational Models, Datasets, Application Development, Startups, Safe and Trusted AI, and Skilling.
Key Details
- Budget: ₹10,000 crore; administered by the Ministry of Electronics and Information Technology (MeitY)
- Compute Pillar: India's common compute capacity has crossed 34,000–38,000 GPUs (publicly available AI cloud)
- Datasets Platform: IndiaAI has over 7,400 datasets and 273 AI models across 20 sectors for researchers and startups
- BharatGen received ₹900 crore under the Mission — the largest beneficiary of the sovereign LLM initiative
- NVIDIA is collaborating with Indian cloud providers (Yotta, L&T, E2E Networks) for AI factory infrastructure
The sovereign models unveiled at the Summit are the direct output of the IndiaAI Mission's Foundational Models pillar — demonstrating that state-supported compute and funding can produce globally competitive AI in less than two years.
Sovereign AI — Concept, Strategic Rationale, and Global Context
Sovereign AI refers to artificial intelligence systems that are developed, trained, deployed, and governed within a country's own jurisdiction, ensuring national control over data, infrastructure, algorithms, and regulatory frameworks. The concept is driven by concerns about: (a) data sovereignty — keeping sensitive citizen/government data within national borders; (b) strategic autonomy — not being dependent on foreign AI systems for critical infrastructure; and (c) cultural alignment — ensuring AI outputs reflect the country's languages, values, and context.
Key Details
- Sovereignty in AI: national control over training data, compute, model weights, and deployment
- India's strategic edge: largest youth population, biggest tech talent pool, 22 scheduled languages + hundreds of dialects
- Sarvam AI's models: trained entirely in India on domestic GPUs using NVIDIA NeMo framework
- BharatGen's Param2: 17B parameter Mixture-of-Experts (MoE) architecture; multilingual and multimodal
- Gnani.ai's Vachana: text-to-speech with voice cloning across 12 Indian languages from under 10 seconds of reference audio
- Global comparators: USA's NAIRR initiative, EU's EuroHPC, France's Mistral AI
The minister's announcement that sovereign AI models are "available at the summit" signals India's transition from a consumer of foreign AI platforms to a producer of its own foundational models — a strategic shift with implications for data governance, national security, and economic competitiveness.
PM Modi's MANAV Framework and India as a Global AI Hub
At the Summit, PM Modi presented India's vision for AI through the acronym MANAV (meaning 'human' in Hindi): Moral and Ethical Systems (AI built on ethical guidelines); Accountable Governance (transparent rules and robust oversight); Natural and Inclusive (AI accessible to all, in local languages); Adaptive and Innovative (flexible to new challenges); Vibrant Ecosystem (collaboration between government, industry, academia). The Summit brought together over 20 heads of state and 500+ AI leaders from 100+ countries, making it the largest AI gathering in the Global South.
Key Details
- Venue: Bharat Mandapam, New Delhi (February 16–21, 2026)
- Participation: 20+ heads of state, 500+ AI leaders, 100+ countries
- First-ever global AI summit hosted by a Global South nation
- India's data centre capacity: 1.5 GW across 7 major cities (end of 2025); projected 1.7 GW by end of 2026
- PM Modi's statement: "We invite the whole world's data to reside in India"
- National Programme for Artificial Intelligence (NPAI) and the broader IndiaAI Mission form the policy spine
The Summit positions India as both a host of global AI diplomacy and a producer of sovereign AI — leveraging its data abundance, talent, and infrastructure to attract global data residency and AI investment.
- IndiaAI Mission budget: ₹10,000 crore (approved March 2024)
- Sovereign AI models unveiled: Sarvam AI (30B + 105B LLMs), BharatGen Param2 (17B MoE), Gnani.ai Vachana TTS
- BharatGen funding: ₹900 crore under IndiaAI Mission
- India's GPU count (IndiaAI compute): 38,000+ existing; +20,000 announced at summit
- Secure strategic GPU cluster: 3,000 next-gen GPUs
- Reliance + Adani AI/data investment pledge: combined $210 billion
- India's data centre capacity: 1.5 GW (end-2025), projected 1.7 GW (end-2026)
- Summit: February 16–21, 2026, Bharat Mandapam — first global AI summit in Global South
- MANAV framework: Moral-Accountable-Natural-Adaptive-Vibrant AI governance vision