Is Capitalism Broken? Why the World Economy Now Depends on a Handful of Companies and Countries

This capitalism is broken; bring us another one. Amid the frenzy of AI, the dismantling of international institutions, mounting global danger and at least two wars that are shaping the international agenda, it is time for a serious conversation about where the world is heading. The founding principles of the capitalist system—competition, freedom of enterprise, the free movement of capital and labor—have today given way to monopolies and oligopolies in key, vital sectors. Freedom of enterprise is being curtailed by regulatory, protectionist and sanctions mechanisms; capital has ceased to be scarce and, in the age of crypto, no longer performs a large part of the functions assigned to it; and labor migration has effectively been halted by the harshest immigration policies in history, with all of this happening against the background of a demographic crisis.

Globalization is often described as a sprawling web of connections in which goods, capital and technology move freely between countries. Yet the most sophisticated production chains have turned out to be acutely vulnerable. They depend on a small number of firms and engineering schools in key industries that have spent decades accumulating knowledge, equipment and the trust of their customers. State control over those chains has now become the foundation of policies whose aims are often incompatible with free enterprise and the public good.

The world economy today rests, de facto, on a few countries and companies on which everyone else depends. The Netherlands supplies unique lithography equipment; Taiwan, a large share of the most advanced chip manufacturing; the United States, data centres, key architectures, software ecosystems and cloud platforms; Japan, critical materials and precision equipment; South Korea, advanced memory; and China, the processing of rare-earth elements, magnets and large segments of clean technology.

De facto monopolies deepen the imbalances that already exist. In some cases, a market is divided between two or three producers. But even an oligopoly of that kind can be fragile if all its members depend on the same materials, the same suppliers or the same political decisions. Competitors, where they emerge at all, cannot offer the required performance, volume, quality and delivery times. This is how bottlenecks arise—and they turn strategic once any disruption spreads far beyond a single market. A halt in supplies of lithography equipment hits not only chipmakers but carmakers, telecommunications manufacturers and arms producers, along with cloud services and medicine. It is precisely this capacity of one local shock to cascade that defines systemic vulnerability. Black swans do not fly in from outside; they live inside the system.

The clearest example is the semiconductor industry. No single country controls the whole cycle, but individual stages of it are extraordinarily concentrated. The Dutch ASML remains the sole supplier of the series-produced extreme ultraviolet (EUV) lithography systems needed for the most advanced logic chips. Its 2025 report records the delivery of 48 EUV systems—a handful of machines on which the world’s most modern chip plants depend.

The next node is Taiwan. TSMC is not a monopoly chipmaker, but its combination of advanced process nodes, scale, product quality and neutral contract model is hard to reproduce.

There are several further layers of concentration around the fabs. Designing complex chips relies on EDA software, where the leading positions are held by Synopsys, Cadence and Siemens (70% of the market). Production equipment is divided among narrow specialists—ASML, Applied Materials, Lam Research, KLA, Tokyo Electron and others. Japanese companies are particularly strong in silicon wafers, photoresists, ultra-pure chemicals, wafer dicing, testing and photomask inspection. There is no single “Japanese monopoly” here, but there is a dense ecosystem of niche leaders who often cannot be replaced quickly.

Artificial intelligence deepens this dependence. An NVIDIA accelerator is valuable not merely as a piece of silicon but as part of the CUDA software ecosystem, its networking solutions and ready-made platforms. Yet the silicon itself has to be fabricated at an advanced plant, packaged together with HBM memory and installed in a data centre. HBM, in turn, comes from a limited group of suppliers—above all SK Hynix, Samsung and Micron. The real bottleneck, then, is not a single corporation but a sequence of interdependent leaders.

Geographical concentration is even more pronounced in rare-earth materials. Deposits exist in many countries, but ore is only the beginning. It has to be concentrated, then the chemically similar elements separated, then metals and alloys produced, and only then high-quality permanent magnets. According to the International Energy Agency, in 2024 China accounted for 60 percent of world mining of magnetic rare-earth elements, 91 percent of their refining and 94 percent of sintered permanent magnet production. It is those last two figures that create strategic dependence. A new mine can be opened, slow and costly though that is; reproducing the full industrial loop from separation to a stable series-produced magnet is far harder.

The export restrictions of 2025 demonstrated the practical dimension of this concentration: the drop in supply forced some carmakers to cut output or halt production temporarily. Alternatives such as Lynas and MP Materials matter strategically, but diversification takes more than new mines. Outside China it is magnet production that will remain the principal bottleneck, and to cover projected demand in full by 2035 the capacity for mining, refining and magnet manufacturing will have to be increased two- to sixfold respectively, over and above the projects already announced.

A shrinking labor force makes robotics critically important. The industrial robot market is an oligopoly. FANUC, Yaskawa, ABB and KUKA have large installed bases but compete with other Japanese, European, American and Chinese manufacturers. A robot, however, is a system of critical components: servo drives, gearboxes, encoders, controllers, sensors, machine vision and software. In some niches the choice of suppliers is very narrow, and swapping a component requires redesign, testing and fresh certification. The same is true of CNC machine tools. Systems from FANUC, Mitsubishi Electric and Siemens are embedded in thousands of factories; their strategic power comes not only from equipment sales but from accumulated compatibility, servicing and staff skills.

In space transport, the most striking concentration has formed around SpaceX. The company holds no legal monopoly on launch: there are Chinese rockets, Ariane 6, Japan’s H3, Vulcan, Electron and other systems. But Falcon 9 has combined reusability, a high flight cadence, substantial available capacity and a satellite programme of its own. The vertical chain running from satellite design to launch and the Starlink service creates a scale that competitors have to catch up with on several levels at once.

Similar logic applies in digital infrastructure. The artificial intelligence industry is controlled by the familiar five trillion-dollar companies, and in the number of data centres the United States is in a class of its own. Cloud computing is concentrated around AWS, Microsoft Azure and Google Cloud. Mobile ecosystems are organized almost entirely around Android and iOS. In frontier AI there are several strong laboratories, so this is not a monopoly. But access to capital, computing infrastructure, data, energy and distribution channels creates exceptionally high barriers to entry. The more basic services migrate to a handful of platforms, the wider the blast radius of a technical failure, a cyberattack or a political restriction.

Concentration of production is not in itself a flaw in the system. It often arises because specialization lowers costs, speeds up innovation and allows expertise to accumulate. Building a second complete chain merely “just in case” is expensive, and in normal times its capacity may sit underused. That is why markets gravitate towards one or a few of the most efficient centres.

The problem lies in the gap between private efficiency and public risk. A company optimizes its costs and its inventories, but does not necessarily reckon with the losses that its shutdown would inflict on dozens of other sectors. A state may support a national producer, yet in doing so intensify trade conflicts, subsidy races and the fragmentation of technical standards.

Uncertainty grows all the more because every node is connected to the others. A new data centre requires advanced chips, HBM memory, networking equipment, electricity and cooling. A robotized factory depends on magnets, power electronics, sensors and software. A space system draws on the same semiconductor and materials chains. In a network like this, a shortage of one component shifts prices, investment plans and political decisions across many industries at once.

The answer cannot be complete national self-sufficiency: modern technologies are too complex and the necessary knowledge too widely distributed. The realistic strategy is to reduce the consequences of any single node failing. That calls for several complementary approaches—geographical diversification, backup suppliers, strategic stockpiles wherever materials can be stored, long-term contracts, interoperability standards, recycling, substitution for scarce materials and the joint financing of new capacity.

Duplicating a plant, however, is not the same as duplicating an ecosystem. That requires engineers, second- and third-tier suppliers, metrology, servicing, software tools and years of production training. Which is why decisions taken today can reduce risk only towards the end of the decade. In the short term, what matters is transparency of supply chains, scenario planning and a readiness to allocate scarce supplies without resorting to chaotic bans.

This interdependence is a source of productivity and of instability at the same time. It makes the world richer in normal times, but it turns local accidents, wars, sanctions, export controls or corporate miscalculations into global shocks. The central economic question of the coming years, therefore, is not how to eliminate every monopoly—that is impossible—but how to keep the specialization that cannot be avoided from hardening into unmanageable fragility.