Jensen Huang's pitch for sovereign AI

For decades, tech luminaries offered one universal piece of advice to young people: learn to code. But speaking at the World Governments Summit in Dubai, NVIDIA CEO Jensen Huang boldly countered that gospel.

In fact, Huang said the exact opposite is now true. NVIDIA's mission is to create technology so intuitive that nobody needs to program anymore. Since the programming language is now simply human, everybody in the world has basically become a programmer.

Beyond flipping computer science dogma on its head, Huang had a much bigger agenda during his fireside chat with the UAE's Minister of AI, Omar Al Olama: selling the idea of Sovereign AI.

Owning your own intelligence

Huang urged leaders from over 150 countries not to let AI mystify them. Instead, every nation must own the production of its own intelligence.

The logic is simple: your data codifies your culture, your society's common sense, and your history. If you don't build models with it, someone else will define it for you. When asked how developing nations should start, Huang claimed building basic infrastructure is neither that costly nor that hard.

The UAE is already sold. The Gulf state is aggressively developing local LLMs, and Moro Hub, a subsidiary of Digital DEWA, just announced a partnership with NVIDIA to build a green data center. With PwC estimating a $320 billion AI boost for the Middle East by 2030, the region is moving fast.

Dubai is just the latest stop on Huang's six-month world tour, following high-level visits to Canada, France, India, Japan, Malaysia, Singapore, and Vietnam.

Why it matters

Jensen Huang might frame Sovereign AI as a noble defense of local heritage, but make no mistake: this is one of the most brilliant enterprise sales pitches we have ever seen.

Big Tech's AI capex can only go so high before hitting fiscal realities. By convincing heads of state that missing out on AI means surrendering their cultural identity and national security, NVIDIA is expanding its addressable market from a handful of Silicon Valley cloud giants to entire nations.

Claiming that building domestic AI infrastructure isn't that costly or difficult feels a bit disingenuous coming from the man selling the hardware. Buying the GPUs is the easy part; the ongoing talent drain, electricity consumption, and algorithmic maintenance will pose massive hurdles for emerging economies.

As for the death of programming, everyday software builders shouldn't throw away their keyboards just yet. Natural language makes prototyping trivial, but complex architecture and deep domain problem-solving still require rigorous technical thinking, even if prompt boxes replace traditional code editors.