AI centralized in tech giants: The era of data colonialism

Colonial powers once competed for land, gold, oil, labour and so on. Today, another resource is driving a new race for power - and billions of us produce it every day: data.

Whenever we search, scroll, click, make a purchase or have a conversation, these all leave behind information about who we are, what we want and how we behave. We hardly notice it. Yet collectively, these tiny digital traces have become a powerful economic resource - one increasingly controlled by a handful of technology giants.

Needless to say, we can barely imagine a day without the internet and social media. AI chatbots have now joined that routine, helping us write, search, create, learn and make decisions. But beneath their convenience lies a larger question: when billions of people continuously generate the raw material for the AI economy, who actually owns it - and who benefits from it?

The giants behind the AI economy

The power of the AI economy does not sit in one place. It is spread across search engines, social media platforms, cloud services, online marketplaces and the computing hardware that connects them. What these companies have in common is that they occupy different points along the same chain: collecting information, processing it, predicting behaviour and providing the infrastructure needed to turn all of that into a business.

Google is a good example of how several parts of that chain can come together. Billions of searches reveal what people are curious about, interested in and, sometimes, ready to buy. YouTube adds another stream of information through what people watch, skip, search for and engage with. Google's AI systems, including Gemini, operate within this much larger ecosystem, giving the company access to vast digital infrastructure and enormous amounts of user interaction.

The value does not come from Google simply selling someone's search history. Instead, the activity of millions of users generates information that can help improve services, personalize experiences and make advertising more effective. In other words, everyday behaviour becomes part of a much larger commercial system. According to Statista, Google generated $294.69 billion in advertising revenue in 2025.

The same dynamic appears on social media, although the type of information is different. On Facebook and Instagram, people constantly signal their interests through likes, comments, shares, follows, searches and viewing habits. Meta can use these signals to recommend content and improve the way advertisements are targeted. The better its systems understand what holds someone's attention, the more valuable that attention becomes to advertisers.

Then there is the infrastructure underneath these platforms. Microsoft, through Azure and its partnership with OpenAI, occupies a different position in the AI economy. Its importance comes not only from data but from the enormous computing and cloud capacity required to train, run and distribute AI systems. Amazon operates across another part of the chain. Every search for a product, purchase, review and comparison on its marketplace produces information that can feed recommendations, demand forecasting, logistics and advertising.

And none of this works without computing power. Nvidia sits further down the chain, supplying the advanced chips and computing systems used to train and operate many of today's AI models. As companies invest more heavily in AI, the demand for these chips and the infrastructure around them grows as well.

Seen together, these companies reveal something important: the AI economy is not simply about who has the most data or the smartest chatbot. Power comes from controlling different pieces of the system. Consumer platforms generate enormous streams of behavioural information. Cloud companies provide the computing infrastructure. AI developers turn data and computing power into models and services. Chipmakers supply the hardware that makes large-scale computation possible.

That is what makes the current concentration of AI power different from a simple race between technology companies. The companies may compete with one another, but they also occupy interconnected positions in an ecosystem where data, infrastructure, software and computing power reinforce each other.

The new colonial relationship

A user in Bangladesh can generate data while using an American or Chinese digital platform. That data can contribute to systems developed elsewhere. The result is that AI products can then be sold back to users, businesses and governments in countries that had little role in building the underlying infrastructure.

The pattern is familiar: data flows outward, technology flows inward, and revenue flows upward. Though that is not to say that every foreign technology company is exploiting developing countries, nor does it mean users are passive victims. Digital platforms provide enormous benefits, often at little or no direct monetary cost.

But there is an undeniable structural imbalance. The countries and corporations that control data, computing power, chips, cloud infrastructure and frontier AI models possess considerably more bargaining power than those that primarily provide users and data.

The World Bank has similarly warned that AI infrastructure and capabilities remain highly concentrated, with developing countries facing disadvantages in computing, connectivity, skills and data.

Information, so to speak, is already a valuable asset. And the more information a country can generate, control and use, the greater its potential economic and technological power. The same logic once applied to lands, minerals, oil, labours and agricultural commodities. At that time the colonial needed ships, armies and territories. Whereas the new powers may need something else like data centres, algorithms, chips and billions of connected humans.

The central struggle of the AI era, therefore, may not simply be about creating smarter machines. It may be about controlling the resources from which machine intelligence is built. And if data is becoming the new oil, the uncomfortable question is no longer whether people are producing the resource. They are. The real question is who gets to control it, who profits from it, and who gets left with little more than the role of supplying it.