As the world descends on New Delhi for India’s AI Impact Summit this week, the questions on the table could not be more consequential.
Artificial intelligence has become a core digital infrastructure. It is in our search results, our inboxes, our workflows. According to analytics and advisory company Gallup, 45% of employees now use AI at work. AI mediates how we find information, interact with services, and work.
The global AI industry is presently dominated by a handful of corporations offering proprietary models. On its face, the value proposition is compelling: A single provider handles the model, hosting, guardrails, and billing. But this dominance stands in the way of widespread economic growth and genuine sovereignty.
Closed models cannot fully accommodate the contextual nuances, languages, and customizations that different societies and cultures require. In a world that has grown more polarized and protectionist — where technology platforms are increasingly enlisted as instruments of state policy — building critical national infrastructure on systems you don’t own, and cannot audit or adapt, is an enormous and growing risk.
The challenge has an economic dimension as much as a political one.
A state concerned with AI sovereignty in 2026 cannot credibly justify financing a foreign, vertically integrated AI stack while neglecting investment in domestic and open-source alternatives.
Investing in hyperscalers can minimize short-term costs but it also entrenches digital rents paid to foreign entities, maximizes long-term dependency on unreliable partners, and dramatically increases exit costs.
If governments finance dependency, dependency is what they will get.
If governments finance dependency, dependency is what they will get.
The current absence of large-scale private funding for open AI infrastructure reflects its public-good characteristics, not its capabilities. Open-source models already routinely achieve 90% or more of the performance of proprietary systems at a fraction of the cost. Investing in open AI frameworks is investing in digital public infrastructure because it yields benefits through lower costs, retained policy autonomy, and economy-wide productivity gains.
It also offers something proprietary systems cannot: democratic legitimacy.
Countries do not capture value by reselling foreign stacks — they capture it by building differentiated products on cheaper, open, shared foundations. This is an industrial policy that promotes competition and the accumulation of domestic capabilities, not a rejection of domestic industry.
Open infrastructure expands the competitive surface for local firms rather than concentrating it in a few foreign providers.
This logic has proved itself before: The internet did not emerge from private actors alone, but from sustained public investment in open technologies. From Linux to Apache, shared open-source foundations have become the backbone of the global digital economy, enabling private innovation while preventing capture at the infrastructure layer. From CERN to Airbus to Galileo, the lesson is consistent: When states co-finance open or shared foundations, private innovation flourishes above them. When they finance access instead, dependence hardens. AI is at exactly that inflection point.
Sovereignty, however, does not mean solitude.
The costs of developing open-source AI can be shared.
As Canadian Prime Minister Mark Carney put it at Davos: “Collective investments in resilience are cheaper than everyone building their own fortress.”
The opportunity is not for each country to construct its own walled garden. It is for nations to collaborate around open standards and shared infrastructure, rejecting the false binary of platform dependency or isolation, and instead building AI futures they actually control.
At Mozilla, we are committing our billion-dollar-plus reserves to open-source AI capability development — investing in existing companies, establishing new ones, funding research and development, building training programs, and mapping open-source AI capabilities and gaps across the stack. We are already in discussions with governments and partners to establish a multi-stakeholder investment and development program that leverages the open-source community to build at real scale and speed.
New Delhi is the right place to make that commitment collectively.
If we want resilient, open, trustworthy AI ecosystems, we must finance them as such — not as an act of idealism, but as a hard-headed investment in sovereignty, resilience, and democratic legitimacy. Countries that want AI sovereignty must be part of this conversation, not as observers, but as co-investors and co-builders. The longer we wait, the harder it gets to change course.
