India AI Impact Summit 2026: Modi calls for human-centric AI, 20,000 GPUs pledged

New Delhi became the global centre of gravity for artificial intelligence this week as the India AI Impact Summit 2026 brought together heads of state, technology executives, and researchers from over 40 nations for five days of deliberation on the future of AI — and India's role in shaping it.

Modi's Human-Centric AI Vision

Prime Minister Narendra Modi, inaugurating the summit on February 16, framed AI not as a tool of the rich but as a lever for the historically marginalised. "AI must be the great equaliser, not the great divider," he said, advocating for the democratisation of AI as a tool for inclusion and empowerment, particularly for the Global South. Modi drew explicit parallels to India's UPI success — a technology that brought 500 million unbanked Indians into the formal financial system — and argued that AI could achieve the same for healthcare, education, and agriculture across Africa and Southeast Asia.

20,000 GPUs and the Compute Race

Union Minister for Electronics and IT, Ashwini Vaishnaw, announced the most concrete commitment of the summit: India will add 20,000 Graphics Processing Units to its existing 38,000 under IndiaAI Mission 2.0. Orders have been placed, and deployment is expected within six months. The move is designed to provide high-quality compute access to startups, researchers, and students — a direct response to criticism that India's AI ambitions were bottlenecked by insufficient infrastructure.

NVIDIA backed the commitment with its own announcement: the company is collaborating with Indian cloud providers — Yotta, L&T, and E2E Networks — to establish advanced AI factories powered by tens of thousands of NVIDIA GPUs, including over 20,000 Blackwell Ultra GPUs for Yotta's Shakti Cloud.

Tata–OpenAI Partnership

In what may be the summit's most commercially significant announcement, the Tata Group and OpenAI revealed a strategic partnership to develop 100 megawatts of AI infrastructure in India, with scalability to 1 gigawatt. The partnership also includes initiatives to boost enterprise AI adoption and expand AI skilling across Tata's workforce of over 900,000 employees. Analysts noted that this represents a shift in OpenAI's strategy — previously focused on the US, Europe, and Japan — toward emerging markets at scale.

Jio's $110 Billion AI Bet

Reliance Jio unveiled its 'Nation-First AI Stack' at the summit, designed for India's unique socio-economic context. The stack emphasises Indian language-based data foundations, a multilingual intelligence layer, and sector-specific application modules. This was accompanied by Jio's broader announcement of a $110 billion (₹10 lakh crore) investment over seven years into AI, with over 120 megawatts of AI-ready data centre capacity expected to come online in the second half of 2026.

BharatGen and PARAM 2

BharatGen, a government-backed consortium of academic institutions, showcased its efforts to build India's native AI models, including PARAM 2 — a multilingual model trained on Indian language data. The consortium announced that PARAM 2 has been benchmarked against leading global models and shows competitive performance in Hindi, Tamil, Telugu, and Bengali text understanding, while using significantly less compute.

What Comes Next

The summit concludes on February 20, with working groups expected to publish recommendations on AI safety, data governance, and compute sovereignty. For India, the question is whether these announcements translate into operational reality — or remain, as critics have argued about previous tech summits, at the MoU stage. The ₹10,372 crore already invested under the IndiaAI Mission provides a baseline, but the true test will come when the 20,000 GPUs are deployed and the first cohort of researchers begins training models on Indian infrastructure.