Artificial intelligence introduces a systemic discontinuity in the relationship between knowledge, economic value, and institutional power. It does not act as a neutral technology that spreads uniformly, but as a factor that reorganizes the geography of cognitive production, redefining the places where knowledge becomes decision and decision becomes value. In this transition, knowledge ceases to be a diffuse resource and tends to concentrate in specific nodes, profoundly altering economic balances and global hierarchies.
The centralization of advanced competencies represents the first structural feature of this transformation. Artificial intelligence systems require a rare combination of computational infrastructures, access to large volumes of data, and strategic integration capabilities. According to OECD data, more than 70 percent of patents related to advanced AI are registered in a limited number of global metropolitan areas, which function as hubs attracting highly qualified human capital and long-term investment. Productive knowledge no longer follows traditional manufacturing or service supply chains, but accumulates in restricted ecosystems, often distant from the places where the material execution of work takes place.
This concentration produces a shift in economic relevance. Economies that fail to secure these cognitive nodes do not disappear, but see their capacity to influence development trajectories reduced. Value is increasingly generated upstream, in the phases of algorithm design, model definition, and control of information flows. Downstream activities, while remaining necessary, become interchangeable and more easily compressible. This dynamic is also observable in the world of multinational enterprises, where strategic functions progressively separate from operational units, transforming the latter into terminals of decisions made elsewhere.
At the same time, new poles of cognitive power emerge that do not necessarily coincide with the borders of traditionally dominant states. Advanced research universities, large technology platforms, high-performance computing centers, and venture capital hubs constitute a transnational network in which knowledge is continuously transformed into decision-making capacity. According to estimates by the McKinsey Global Institute, firms operating in proximity to these ecosystems show productivity growth rates significantly above average, not as a result of economies of scale, but due to their ability to anticipate scenarios and shape emerging markets. In this sense, knowledge is not merely a productive input, but a lever of influence over the rules of the economic game.
The effects on work are profound and ambivalent. On the one hand, the concentration of competencies accelerates selective migration flows toward these centers, reinforcing polarization between central and peripheral territories. On the other hand, AI enables a certain degree of virtual mobility, allowing part of cognitive work to be performed remotely. However, this openness does not compensate for the loss of centrality of local contexts that do not participate in the production of strategic knowledge. Mobility becomes asymmetric: it favors those who already possess cultural capital and access to global networks, while leaving behind those anchored to economies in cognitive decline.
In this context, the link between knowledge and power becomes particularly evident. Michel Foucault showed how knowledge is never separable from the dispositifs that make it operative. Artificial intelligence can be read as one of these dispositifs, capable of organizing what is visible, calculable, and therefore governable. The concentration of productive knowledge is not an accidental outcome of innovation, but a form of governance that redefines the possibilities of action for firms, workers, and institutions. Those who control models and computational infrastructures exercise a power that precedes formal political decision-making.
For firms, this transformation translates into a challenge of leadership and strategic positioning. Investing in technology is not sufficient if it is not accompanied by an understanding of one’s position within the new cognitive cartography. Organizations that fail to secure an autonomous decision-making center risk becoming efficient executors of strategies defined elsewhere, progressively losing their capacity for vision. Conversely, firms that grasp the geographical dimension of knowledge are able to build a competitive advantage that goes beyond the short term, influencing standards, practices, and market expectations.
The algorithmic mediation of decision-making processes further amplifies cognitive limits rather than eliminating them. The concentration of knowledge can generate an illusion of control, while in reality increasing systemic complexity and the distance between those who design models and those who bear their effects. Responsibility thus shifts to the architectural level: choosing how to organize information flows, how to distribute competencies, and how to prevent the separation between center and periphery from becoming an irreversible fracture.
The new cartography of value emerging from artificial intelligence is not only economic, but social and institutional. It redefines the relationship between work and knowledge, between mobility and belonging, between geographic proximity and decision-making capacity. It does not offer predetermined outcomes, but opens a space of tensions in which concentration and dependence coexist. Within this space, the challenge for states and firms is not to halt the movement, but to orient it, recognizing that every concentration of cognitive power generates a responsibility proportional to its systemic influence.
Global AI Observatory