The contemporary economy is undergoing a structural transformation in the nature of capital and in the mechanisms of its accumulation. In data intensive economic systems, artificial intelligence operates as an enabling infrastructure of a new form of economic power, less visible but more persistent. Capital no longer grows solely through the expansion of productive capacity or organizational efficiency, but through the progressive sedimentation of data, predictive models and analytical capabilities that orient future decisions even before they are formulated. This shift profoundly alters the relationship between value, labor and institutions.

Data do not constitute a resource comparable to traditional raw materials. Unlike them, they are not consumed through use, but increase in value through repetition and combination. Every economic interaction generates information that feeds machine learning systems, improving their performance and expanding their scope of application. According to OECD estimates, more than 90 percent of the data available today has been produced in the last few years, and a growing share of this data is used to train models capable of anticipating consumption behavior, logistical flows and market dynamics. Data thus act as reflective capital, capable not only of describing economic reality, but of modifying it through the orientation of decisions.

This transformation introduces accumulation mechanisms radically different from those of industrial capitalism. Investment no longer translates primarily into new plants or increased labor force, but into the strengthening of algorithmic learning cycles that generate cumulative returns. Better data generate more accurate models, which attract new users and new transactions, producing additional data. The process is circular and self reinforcing. Analyses by the International Monetary Fund show that data intensive firms record increasing marginal returns, with concentration effects that exceed those observed in traditional industrial sectors. Accumulation takes on an exponential form, structurally widening the distance between those who lead the process and those who follow it.

In this context, barriers to entry change in nature. They are no longer only financial or regulatory, but technological and temporal. Access to large historical data repositories, the availability of adequate computational infrastructures and the time required for systems to learn become indispensable requirements. Time itself is transformed into a competitive factor. Late entrants do not simply lose market share, but find themselves excluded from entire value ecosystems. According to the European Commission, in advanced digital markets the advantage of first movers tends to consolidate over the long term, making the emergence of new independent actors increasingly difficult.

For firms, this configuration redefines the concept of strategy. Leadership no longer coincides solely with entrepreneurial vision or operational execution capability, but with the conscious management of informational assets that grow silently within daily activities. Deciding which data to collect, how to integrate them and how to translate them into predictive capacity means influencing the future trajectories of the organization. This represents a form of responsibility less visible than traditional investments, but more pervasive, because it links present choices to long term effects that are difficult to reverse.

At the macroeconomic level, data intensive accumulation tends to produce increasingly concentrated market structures. Growth through acquisitions or explicit anticompetitive practices is not necessary. Increasing returns linked to data and models favor a few actors capable of sustaining continuous investment in digital infrastructures. Capital becomes dematerialized and opaque, escaping classical economic metrics. Indicators such as physical capital or labor intensity become increasingly inadequate to describe the distribution of economic power. Reports by the World Bank highlight how the share of value attributable to intangible assets, data, software, algorithms, now exceeds that of material assets in many advanced economies.

This opacity recalls the analyses of Fernand Braudel on the deep and structural dimension of capitalism. Braudel observed that the most durable forms of economic power did not manifest themselves in visible markets, but in the slow and hidden layers of the economy. In a similar way, data based accumulation operates beneath the surface of daily transactions, building advantages that emerge only in the long term and ultimately determine the overall architecture of the economic system.

The social and ethical implications of this process are significant. When capital accumulates through invisible and hardly accessible resources, the asymmetry grows between those who can orient the future and those who are forced to adapt to already defined models. Economic autonomy becomes an indirect function of access to data and informational infrastructures. Workers, firms and even public institutions may find themselves dependent on systems they do not fully control, accepting decisions generated by opaque logics in the name of efficiency.

In the long term, data intensive economic systems pose a question of institutional sustainability. A model of accumulation founded on informational concentration can generate growth without a corresponding diffusion of value, fueling social tensions and systemic fragilities. For entrepreneurs and public decision makers, the challenge does not consist in restraining data accumulation, but in making its logics intelligible and its effects governable. Capital in the algorithmic economy is not only what is owned, but what is made transparent, accountable and integrable into an economic order capable of maintaining cohesion.

Capital accumulation in the data economy occurs less and less in the visible places of production and increasingly in the silence of informational infrastructures. In this discreet but decisive space, the conditions of competition, value distribution and economic legitimacy are redefined, outlining a transformation of capitalism that proceeds without noise, but with structural effects destined to endure.

Global AI Observatory