Top takeaways from The 2026 India AI Impact Summit

The India AI Impact Summit 2026 concluded on February 20 at Bharat Mandapam, marking a decisive pivot for the Global South. While the "handshake snub" between Sam Altman and Dario Amodei dominated social media, the actual outcomes of the summit represent a massive restructuring of the global AI supply chain.

With over $250 billion in infrastructure pledges and the participation of 118 countries, the event shifted the conversation from theoretical "AI Safety" to industrial-scale "AI Sovereignity."

Here are the biggest highlights from the event.

The $230 billion infrastructure "arms race"

The summit was defined by staggering capital commitments from both Indian conglomerates and American tech giants, aimed at turning India into a global compute hub.

• Reliance’s $119B "Jio Brain": Mukesh Ambani announced a 10 lakh crore rupee ($119 billion) investment over seven years to build a national AI infrastructure, including green-energy powered gigawatt-scale data centers in Gujarat.

• Adani’s 5GW Target: The Adani Group pledged $100 billion to expand its data center capacity to 5 gigawatts, specifically optimized for high-density AI workloads.

• Microsoft & Google: Microsoft reiterated a $50 billion "Global South" investment plan, while Google unveiled the "America-India Connect" subsea cable project and a new $15 billion AI Hub in Visakhapatnam (Vizag).

• The "MANAV" vision: A new governance framework

Prime Minister Modi introduced the MANAV (Human) framework as a counter-proposal to Western and Chinese regulatory models. The vision emphasizes five pillars:

• Moral and ethical systems

• Accountable governance

• National sovereignty (data ownership)

• Accessible and inclusive technology

• Valid and legitimate lawful systems.

The rise of "Bharat" models

A key theme was the rejection of a "one-model-fits-all" approach dominated by English-centric LLMs.

BharatGen: The government-backed initiative launched Param 2, a 17-billion-parameter model supporting all 22 scheduled Indian languages.

Sarvam AI: The startup revealed its 105B parameter model, claiming it outperforms GPT-4 in regional linguistic nuance and "frugal" token efficiency.

OpenAI’s anchor tenant deal: In a major win for Indian IT, TCS (Tata Consultancy Services) announced that OpenAI will be the first anchor tenant for its Hypervault data center, starting with 100MW of capacity to power "OpenAI for India."

Democratizing the "GPU stack"

To prevent a digital divide, the Indian Ministry of Electronics (MeitY) announced a massive expansion of the IndiaAI Mission:

GPU subsidies: The government is expanding its national GPU pool by 20,000 units (on top of the current 38,000), making high-end compute available to startups for as little as $1 per hour.

Google and the Indian government also partnered to train 20 million public servants in AI literacy, ensuring that the "last mile" of government service delivery is AI-augmented.

"The future of AI cannot be decided by a handful of countries or left to the whims of a few billionaires," noted UN Secretary-General António Guterres during the plenary session, echoing the summit's push for a more multipolar AI world.