India hosts the AI Impact Summit 2026

Background

The AI Impact Summit 2026 was held in New Delhi from 16 to 21 February 2026, at the Bharat Mandapam convention complex ^[1]. It was the first time that the global AI summit series took place in the Global South ^[1]. The meeting ran under the theme "Sarvajan Hitaya, Sarvajan Sukhaya", a phrase meaning welfare for all and happiness of all ^[1].

The published agenda scheduled the adoption of the Leaders' Declaration for 20 February 2026, the Friday of the summit week ^[2].

What happened

More than 100 countries took part, among them 22 heads of state or government, alongside around 10 international organisations ^[1]. The summit website listed Dario Amodei, chief executive of Anthropic, among the key attendees ^[2]. The Indian government said the event drew approximately 600,000 attendees in person ^[1].

Two documents came out of the week. The India AI Impact Summit Declaration was endorsed by 92 countries and international organisations ^[1]. Separately, the New Delhi Frontier AI Impact Commitments were signed by 13 leading model providers ^[1]. The two tracks are distinct: the declaration is a statement endorsed by governments and organisations, while the commitments are pledges made by companies that build frontier models.

On the economic side, investment commitments announced across the AI value chain were expected to exceed US$250 billion ^[1]. That figure covers the chain as a whole, from the hardware and data-centre layers through to the services built on top of models, rather than a single segment.

Key details

The choice of venue and host is part of the substance of the meeting. The summit series had previously been convened in the Global South's absence, and the 2026 edition moved the discussion to a country that is neither a major model developer nor a traditional venue for such gatherings ^[1]. The theme chosen by the host frames AI in terms of broad public welfare rather than competition between developers ^[1].

The attendance numbers give a sense of scale. A stated 600,000 in-person attendees is far larger than a typical leaders' meeting, and the presence of 22 heads of state or government alongside roughly 10 international organisations indicates that the event mixed diplomatic negotiation with a public-facing programme ^[1]. The presence of a frontier-model chief executive among the listed key attendees indicates that the companies building the largest models were represented at senior level ^[2].

Why it matters

The summit produced three concrete outputs that can be tracked afterwards: a declaration endorsed by 92 countries and international organisations, a set of commitments signed by 13 model providers, and a headline figure of more than US$250 billion in announced investment commitments across the AI value chain ^[1]. Each of these is a commitment rather than an implemented measure, and the value of the meeting will depend on what signatories do next.

The location matters for a different reason. Bringing the series to the Global South for the first time put questions of access, cost and public benefit at the centre of an agenda that had largely been set by the countries where the largest models are built ^[1]. The theme of welfare for all is a statement of that emphasis ^[1].

The separate signing by 13 model providers also matters structurally. It creates a record of which companies were willing to attach their names to frontier AI commitments at a multilateral meeting, in a way that can be compared with later meetings and with the companies' own published policies ^[1].

What the summit did not do is settle anything binding. The declaration, the commitments and the investment figures are all announcements; none of the facts around the meeting establish enforcement or review mechanisms. The event is best understood as a point where a large number of governments, organisations and companies put their positions on record in one place, in the Global South, over six days ^[1].