“As nations race to build and govern AI, the question is no longer just who owns the technology, but whose worldview defines the digital future”. - Dhruv Suri
When India hosted the AI Impact Summit in New Delhi last month, it was making a claim that goes well beyond technology policy. The large-scale AI models that now underpin global digital life are not merely computational tools – they are cultural infrastructure, encoding specific languages, worldviews, and assumptions about knowledge itself. Ninety-one countries endorsed the resulting New Delhi Declaration, signalling a growing consensus that who owns and governs these models is a question of global equity, not just market competition.
The governance stakes of this moment are underscored by the hard security dimension of advanced AI: credible reports suggest US and Israeli forces used AI-assisted systems to identify and prioritise targets in strikes on Iran, demonstrating how questions of who owns and governs foundational models have become matters of life and death, not just economic policy. This convergence of diplomatic and military dimensions is precisely what gives gatherings like the New Delhi Summit their urgency.
At the heart of the summit's diplomatic outcome was the New Delhi Declaration on AI Impact, endorsed by 91 countries and international organisations within days of its adoption. Guided by Indian philosophical tenets such as “Sarvajan Hitaya, Sarvajan Sukhaya” (welfare for all, happiness for all) and “Vasudhaiva Kutumbakam” (the world is one family), the declaration emphasises that the benefits of AI should be equitably shared and that access to AI capabilities must not be confined to a narrow group of countries or firms.
The declaration is deliberately broad and voluntary, organised around seven pillars that range from democratising AI resources to building resilient infrastructure. To operationalise these priorities, the summit announced several voluntary initiatives: a Charter for the Democratic Diffusion of AI, a Global AI Impact Commons, and a Trusted AI Commons.
Critics will note the declaration's voluntary character as a structural weakness: in the absence of enforcement, its commitments risk remaining aspirational. That tension is real, but it is also inherent to any multilateral consensus that includes both advanced AI powers and developing nations - the alternative to a voluntary framework is often no framework at all.
Beyond Governance: Foundational Model Diplomacy
The New Delhi Declaration also points to a deeper shift in how states think about technology: the rise of what this article terms “foundational model diplomacy.” Unlike traditional technology diplomacy - once focused on physical infrastructure like ports, telecom networks, or data centres - the geopolitics of AI now extends into culture, knowledge, and identity. Foundational AI models are not merely computational artefacts; trained on massive datasets under particular assumptions, they embed specific worldviews, linguistic patterns, and cultural references that shape how information is represented and which voices are amplified.
Seen this way, AI data centres, compute stacks and, above all, large-scale models are becoming a new kind of cultural infrastructure. They mediate which languages are usable, how knowledge is organised, and how societies understand concepts such as truth, agency, and creativity. This marks a departure from the earlier cloud-computing era, when storage and compute were often treated as fungible commodities and their cultural imprint attracted less attention. Today's foundational models are born of specific datasets, trained under particular assumptions, and influenced by the normative preferences - explicit or implicit - of their creators, echoing older struggles over media, language, and representation in a more algorithmic form.
This “cultural infrastructure” perspective carries profound diplomatic implications. States increasingly recognise that control over models, data governance regimes, compute infrastructure, and technical standards translates into influence over global knowledge ecosystems. Foundational model diplomacy, therefore, must address not only access to technology but also the terms of representation - who defines fairness, which languages and communities are prioritised, and how societal values are encoded in technical systems. India's emphasis on democratic diffusion and inclusive access at the New Delhi Summit reflects this understanding, reframing AI cooperation as a question of social equity and cultural participation, rather than a narrow contest over market share.
Why Foundational Model Diplomacy Matters
In a geopolitical environment marked by superpower competition and technological nationalism, foundational model diplomacy offers a complementary paradigm to both strict regulatory control and laissez-faire market dominance. It starts from three basic insights:
AI capability and influence are entwined: countries and corporations that control advanced foundational models exercise outsized sway over global technical standards, innovation pathways, and even policy debates, because their systems become de facto reference points.
Cultural plurality requires structural support: without deliberate investment in datasets, models, and tools that reflect diverse languages and perspectives, AI systems will tend to reproduce existing imbalances across regions, languages, and social groups, marginalising low-resource communities.
Cooperation must balance competition and shared benefits: models and compute infrastructure cannot be locked indefinitely behind unilateral control if AI is to function as a global enabler; yet cooperative frameworks must still ensure safety, interoperability, and trustworthiness.
By elevating these considerations, the New Delhi Declaration helps shift the narrative from AI as a pure engine of economic growth to AI as a shared social and cultural commons. In this emerging order, data, models, and compute are not only economic inputs but also levers that shape how societies remember, reason, and imagine their futures.
A Way Forward for India
As one of the key convenors of the AI governance conversation, India is well-placed to translate the summit's commitments into sustained global leadership. Doing so will require aligning domestic capacity, foreign policy, and multistakeholder engagement around a coherent agenda for foundational model diplomacy.
Invest in sovereign and inclusive AI infrastructure: India can deepen its AI compute and research ecosystem while ensuring that datasets and models reflect the country's linguistic, cultural, and social diversity. This includes supporting indigenous foundational models, expanding language-first AI research, and strengthening open-source AI platforms and public digital infrastructure that lower entry barriers for researchers and startups.
Champion ethical and inclusive standards: Alongside technical investments, India can work through multilateral and multi-stakeholder forums to develop interoperable standards for explainability, fairness, safety, and accountability that are sensitive both to global norms and to regional realities. Aligning these standards with the New Delhi Declaration's focus on trustworthiness and human-centric development will help build a more consistent global baseline.
Forge strategic partnerships without dependency: Supply-chain collaborations, compute partnerships, and AI research alliances will remain essential, but they should be structured to enhance India's strategic autonomy rather than deepen dependence on any single actor. This means diversifying technology partners, co-developing critical infrastructure, and negotiating arrangements that preserve room for domestic innovation and regulation.
Empower Global South innovation: By positioning itself as a bridge between advanced economies and the wider Global South, India can support capacity-building, knowledge exchange, and equitable access to AI resources for countries with limited technical or financial means. This might involve regional AI centres of excellence, shared compute facilities, or collaborative projects applying AI to public health, agriculture, climate resilience, and digital public services.
Institutionalise foundational model diplomacy: India can institutionalise these efforts by creating sustained diplomatic platforms and technical networks dedicated to foundational AI cooperation. Regular dialogues that bring together states, civil society, academia, and industry around shared norms, infrastructure commitments, and capacity-building roadmaps would prevent the momentum of the New Delhi Summit from dissipating.
Conclusion
The New Delhi Declaration is more than a diplomatic communiqué - it reflects an emerging global AI order centred on equitable access, cultural representation, and shared technological stewardship.
For India, the challenge now lies in converting a summit consensus into durable institutions - ones that ensure the next generation of foundational models reflects not just the worldview of Silicon Valley or Beijing, but of New Delhi, Lagos, and São Paulo too.