The AI Impact Summit 2026 in New Delhi is signaling something far more consequential than a commitment to India’s “inclusive technology.” It’s signaling scale.
The government’s IndiaAI Mission carries an allocation of roughly $1.25 billion to expand domestic AI capacity.
Officials state that more than 38,000 GPUs have been onboarded for shared national access, 12 indigenous foundation models are in development, and more than 30 India-specific AI applications have been approved.
Those are measurable industrial inputs, not the typical abstract policy aspirations to which investors have become depressingly accustomed.
With more than 45 countries represented and leaders including Narendra Modi, Emmanuel Macron, Luiz Inácio Lula da Silva and António Guterres in attendance, the summit frames AI through the lens of responsibility and equity.
Investors should, in my opinion, focus instead on the underlying capital formation.
The first phase of the AI boom concentrated returns in US mega-caps that controlled model training, cloud architecture and enterprise distribution.
This concentration was logical: AI economics reward scale, infrastructure density, and data pooling.
The next phase depends on which countries build domestic processing depth and reduce exposure to external chokepoints.
With this summit, it’s clear that India is attempting to do exactly that.
Aggregating tens of thousands of GPUs within a shared framework creates a national base layer for model training and deployment. Developing indigenous foundation models reduces reliance on foreign architectures. Embedding AI across agriculture, healthcare, education and public services increases domestic demand pull.
Each element reinforces the other.
Participation from executives at Microsoft, IBM, Zoom, Kyndryl and Adobe underscores that global firms recognise where growth markets are consolidating.
India offers scale, policy clarity and expanding digital infrastructure. For multinationals, proximity to such a market reduces time lags, compliance friction and operational risk.
There’s also a geopolitical dimension. AI supply chains remain sensitive to export controls and semiconductor concentration.
Expanding national capacity alters bargaining power. If India shapes AI standards for emerging markets, particularly across the Global South, it influences procurement frameworks and regulatory norms across economies representing billions of users.
Capital markets often fixate on quarterly earnings concentration in a handful of US names. Structural infrastructure buildouts in large emerging economies receive less attention.
But history teaches us that when nations commit capital, hardware acquisition, and talent development simultaneously, they are laying foundations for real growth.
It’s clear that AI is moving from a unipolar to a multipolar phase.
The US retains platform dominance and China retains vertically integrated state coordination.
But India is building scale through infrastructure aggregation and policy-backed expansion.
The critical question for investors is where future leverage accumulates.
The summit’s core message is straightforward: India intends to build its own AI architecture.
Investors around the world will see this as a structural reallocation of AI capacity towards the Global South.
Nigel Green is the deVere Group CEO and founder.
The geopolitical realities of the day make entry into capital-intensive AI business very risky unless one owns the entire stack, beginning with the semiconductor chips that run the AI hardware. India does not have any semiconductor manufacturing capacity. Its investment, if any, will be subject to later “sanctions” (and thus totally dependent on the whims of whoever can sanction) if it invests any money in the field. Talk is cheap. To deveop the country, India should invest heavily, like the Soviets did in their early days, with basic sciences (physics, chemistry, mathematics), metallurgy, chemical engineering, materials science, etc and then build its own national expertise. They think by providing labourers to OpenAI and Microsoft they can compete with those companies, which is nonsense. India should not, at this time, even try to be anything in AI. For one thing, AI is simply brute force application of well-known mathematical techniques. It may or may not help in improving productivity at the margins. If India is doling out rice/wheat to 800 million people at 5Kg per person per month, it is preposterous to think it can compete with anyone in the AI field. Moreover, at this time, all investment in AI is highly speculative and extremely uncertain because a different model may knock off established models and all investments may go up in flames. Is India ready for such a loss? Learn to walk before trying to run. Modi is not even college-educated. Such people, like marketers, will speak of a lot of things of which they have no clue. Proclamations do not make a country the “power” in any field.