Nvidia CEO Jensen Huang urges countries to build their own sovereign AI

Jensen Huang has a pitch for every government on Earth: stop renting your intelligence and start building it at home.

The Nvidia CEO says he will help fund countries that want to develop their own AI, according to Axios. It’s a generous offer from a man whose company sells the chips those countries would need.

What Huang is offering

The idea is called “sovereign AI.” Think of it like a national power grid, but for machine intelligence. Each country owns the data, the models, and the computing hardware instead of plugging into a foreign provider.

Huang has championed the concept since February 2024. His argument is that AI is national infrastructure, comparable to roads or electricity. A country that depends entirely on outside providers for it hands over a degree of control it may not get back.

His sharpest framing is about data. Huang has warned against the risk of exporting data in order to import intelligence. Put simply: a country ships its information abroad, then pays to get insights back from systems it doesn’t own.

On October 15, 2026, during a European tour, Huang pledged an additional $2 billion for Europe’s AI ecosystem. He separately committed $250 million in UK computing resources for AI safety over five years.

The UK commitment is aimed at AI safety and governance research at institutions including Oxford and UCL.

A global buildout already underway

Sovereign AI initiatives are active in more than 20 countries.

Large GPU deployments have been reported in South Korea, Japan, Saudi Arabia, and the UAE. GPUs, or graphics processing units, are the specialized chips that train and run modern AI systems. Nvidia dominates that market.

South Korea stands out. The country is set to receive up to 260,000 Nvidia GPUs, with that delivery tied to late 2025.

Huang has also floated a condition on all this enthusiasm. Nvidia might withhold chips from developers it considers irresponsible, particularly those that don’t meet necessary safety standards.

Why this matters beyond Nvidia’s sales

Governments increasingly want independent technology infrastructure to reduce their reliance on established American tech giants. AI, with its appetite for sensitive data and its potential military and economic weight, is accelerating that instinct.

There’s an irony here worth noting. Countries pursuing independence from US tech companies are often doing it by buying hardware from a US tech company. Sovereignty, it turns out, still runs on Nvidia silicon.

Analyst estimates suggest sovereign AI may contribute approximately 14% to Nvidia’s business, potentially reaching $50 billion annually long-term. A business concentrated among a small number of giant buyers carries risk if those buyers slow their spending. Government demand driven by national security and economic policy may behave differently from corporate capital expenditure cycles.

By putting money into European ecosystems and UK safety research, Nvidia positions itself as a partner in national ambitions rather than simply a vendor.

If Nvidia does follow through on restricting chips to developers that fail safety standards, it raises questions about who sets those standards and how they’re enforced.