AI independence is a myth. Here’s why

A version of this article was originally published in Quartz’s AI & Tech newsletter. Sign up here to receive the latest news, analysis and insights on AI and technology straight to your inbox.

A trillion-dollar bet is appealing to governments around the world, based on the promise of digital independence. Build your own data centers, process your citizens’ data nationally, and capture the economic value of the AI ​​boom instead of watching it flow exclusively to Silicon Valley. From Southeast Asia to the Middle East to Europe, governments are spending billions on what they call “sovereign AI,” a new form of digital independence in the age of intelligent machines.

It is to be presented as digital decolonization. Just as colonial powers once extracted oil and minerals while providing minimal benefits to local populations, a handful of U.S. tech giants are now extracting data, turning it into intelligence, and selling it back to the countries where it came from. Sovereign AI promises to break this cycle.

But there is a fundamental contradiction at the heart of this giant construction. Countries seek their sovereignty by making themselves more dependent on the foreign companies against which they claim to protect themselves.

The paradox of sovereignty

The model plays out everywhere. In Indonesia, the country’s second largest telecommunications provider plans to increase the capacity of its AI server 25 times by 2028 – an investment of around $5 billion that it couldn’t make without Nvidia chips and U.S. cloud partnerships.

Saudi Arabia won preliminary approval to buy 18,000 Nvidia chips during President Donald Trump’s May visit and hopes buy several hundred thousand of the company’s best processors over the next five years.

India is receive an investment of 15 billion dollars from Google. Thailand receives $5 billion from Amazon Web Services. France is building what it calls Europe’s largest AI data campus via Mistral, its national AI leader – in partnership with Nvidia.

Justifications for AI sovereignty vary. Some countries want to ensure that AI models incorporate local languages ​​and values. Others worry about sensitive health data flowing through foreign systems. Small countries argue that national infrastructure ensures cheaper and more reliable access to local businesses and researchers who would otherwise “always last in line” for IT resources.

But none of these projects provide anything close to self-sufficiency. Nvidia, based in the heart of Silicon Valley, dominates the AI ​​chip market with a share of around 90%. Its only serious rival, AMD, is also American. The servers that host these chips mostly come from Dell and Supermicro, both American. Even China, which has built something close to a self-sufficient AI stack, has yet to develop an alternative to American processors.

U.S. export controls on advanced AI chips give Washington leverage over any countries seeking to build AI infrastructure. Saudi Arabia’s chip purchases are still awaiting final approval several months after Trump’s visit. Future sales could be tied to individual trade deals, and countries would find themselves caught between U.S. security requirements and their existing relationships with Chinese technology companies.

The cheaper alternative that no one wants to admit exists

Amid giant spending and already tight budgets, another complication arises: Tech giants could provide the same computing capacity at a lower cost. Amazon and Microsoft already offers a “sovereign cloud” offers with enhanced data controls and dedicated local infrastructure. These products promise data localization and security controls, allowing countries to build national AI models on top of existing infrastructure.

The size of hyperscalers gives them negotiating power with suppliers like Nvidia that some countries lack. They can deliver IT capacity faster and more cheaply than governments building from scratch.

But these commercial solutions cannot guarantee true sovereignty either. American companies remain subject to American jurisdiction, in particular the CLOUD Act, which allows American authorities to access the data that these companies hold anywhere in the world. When French lawmakers pressed Microsoft As for whether it could guarantee that French citizens’ data would not be transmitted to the United States without French approval, the company could not make that promise.

This creates what critics call “sovereignty as a service” countries having the appearance of control while foreign corporations retain real power. This arrangement resembles the build-operate-transfer programs of the colonial era, in which European companies built critical infrastructure in colonies under the guise of development.

THE the modern version may be worse. These public-private partnerships operate with a minimum of transparency. Contracts with tech giants often hide crucial details about data access and algorithmic control from legislative oversight. Subscription-based models give companies the ability to update or deactivate their systems remotely, meaning countries risk having their “sovereign” infrastructure disabled if they do not respect companies’ interests.

None of this means that countries should not invest in AI capabilities. But the gap between sovereign branding and effective control is widening. Countries that have achieved true digital independence have invested in national technical expertise, adopted open standards, and strengthened the ability to switch providers.

Without these elements, national AI projects risk becoming expensive showcases for foreign technology – rather than true infrastructure under local control.