Sovereign AI: the new race for technological autonomy

This article has been translated using artificial intelligence

The Future Trends Forum of the Fundación Innovación Bankinter has brought together international leaders from the scientific, business, and technology sectors to explore one of the most disruptive frontiers of artificial intelligence: Embodied AI. This new generation of AI, which moves beyond the purely digital realm to integrate with the physical world, signals a radical shift in how we interact with technology. From autonomous robots to smart devices that engage with their surroundings, the forum has examined the advances, ethical challenges, and strategic opportunities that define this revolution. It’s a transformation that affects not only industrial and healthcare sectors but also geopolitical balance and technological sovereignty.

Within this context, Jordan Sun—entrepreneur, former military officer, and former Chief Innovation Officer of San Jose, California—issued a stark warning: we are entering a new race for technological power. And those who fail to act in time will find themselves subordinated.

“It’s not about who has the best AI, but who can avoid being subjugated.”

A new geopolitics of artificial intelligence

Sun draws a comparison between the development of AI and the nuclear arms race: today, only four players dominate the foundational AI models, and three companies control the global market for cloud services. This extreme concentration poses a threat not only to the sovereignty of countries but also to the strategic autonomy of businesses.

The dependence on technologies that are neither understood nor controlled places governments and organizations in a vulnerable position. And the solution isn’t about competing on speed—it’s about building local capabilities aligned with national interests.

A lesson from history (and science fiction)

To illustrate this dynamic, Jordan Sun refers to Isaac Asimov’s Foundation universe, where an advanced civilization offers technology to others without allowing them to understand or replicate it. The same is happening today: many countries use AI models or cloud infrastructure without having sovereignty over them.

The historical parallel is also clear. Japan modernized itself after confronting Western powers; China launched its Made in China 2025 plan to reduce its dependence on the U.S. Sun sums it up: “There’s no sovereignty without technological capability.”

Governments and businesses: co-responsible for change

Sun argues that governments must take a more active role by developing policies that promote not just adoption, but also understanding and control of emerging technologies. He points to the U.S. E-Government Act, which institutionalized the role of CIOs in public administration, and to California’s pioneering role in digital innovation policy. In Europe, he cites the GDPR as a robust regulatory model, and in China, measures like the ban on exporting genetic samples to preserve sovereignty over biometric data.

He also highlights the role of the European Union with GDPR and data regulation, or China’s restrictions on genetic data exports. The point is clear: having talent or startups isn’t enough; institutional strategy is essential.

In the private sector, he cited Nuro, a self-driving vehicle startup that has actively collaborated with regulators since its inception. Integrating innovation with the public interest can be a competitive advantage and a way to legitimize technological advancement.

David Shrier, another FTF speaker, reinforces this approach from an academic and policy perspective. In his talk, Shrier analyzes how governments are crafting sovereign AI strategies to reduce dependence on big tech companies and build models that reflect their languages, values, and national priorities. Shrier notes that many governments want to lead in AI without realizing that it’s not just about developing models: you must control the entire tech stack, from chips to energy infrastructure.

“Everyone wants their own LLM, but they ignore the fact that you need talent, hardware, data, and energy. Like energy: no one is totally autonomous.”

You can read his insights in this article:
Sovereign AI: the new global strategy to avoid dependence on Big Tech

Who’s doing it right?

During the Q&A, a key question emerged: which countries are approaching this race for sovereign AI in a smart way? Sun’s answer was nuanced: there are no perfect solutions, but there are promising approaches:

• Israel has created an agile ecosystem that uses other markets as a platform to scale.
• The United Kingdom leads in per capita AI productivity and has heavily invested in fundamental research.
• Singapore, though dependent on external data centers, has built a highly effective narrative around digital sovereignty.
• Abu Dhabi has surprised the world with its AI university and homegrown models like Falcon.
• China is investing massively, although ideological restrictions may limit the development of powerful models in the long term.

The European case is also mentioned: models like Mistral could take off if they gain coordinated political and financial backing.

A global AI governance framework?

Sun concluded his talk by proposing an ambitious idea: a global governance structure for AI models to counterbalance the dominance of a few actors. A sort of technological UN to ensure more equitable, transparent, and secure access to the tools that will define the future.

“It’s inevitable that one country will lead. But we still have time to design a global architecture that mitigates these imbalances.”

For Sun, the race is not just about innovation—it’s about autonomy. In this new map, those who don’t control their technologies will be controlled by them.