Unregulated capitalism has always tended to monopoly. But Big Tech represents a challenge antitrust tools can’t tame.

Scientia potentia est—knowledge is power. The old adage has acquired a sinister connotation with the alarming dominance of Big Tech in the economy and society as a whole. Corporate Europe Observatory recently revealed that the sector is now by far the leading business lobbyist of European Union institutions.
But this is only the tip of the Iceberg of what the Italian economist Ugo Pagano calls ‘intellectual monopoly capitalism’. Knowledge, which should be a (non-rival, non-exclusive) public good, has been privately appropriated by top companies as capital: the share of intangible assets among S&P 500 corporations increased from 17 per cent in 1975 to 90 per cent in 2020.
For Pagano, the dramatic expansion of intellectual-property rights ‘involves the creation of a legal monopoly that can be potentially extended to the entire global economy’. His claim against a strict IP regime echoes the traditional position of economists treating knowledge as a gratuity. Friedrich Hayek, for example, contended:
The growth of knowledge is of such special importance because, while the material resources will always remain scarce and will have to be reserved for limited purposes, the uses of new knowledge (where we do not make them artificially scarce by patents of monopoly) are unrestricted. Knowledge, once achieved, becomes gratuitously available for the benefit of all.
Recent calls for a patent waiver on Covid-19 vaccines graphically illustrate this broader principle: general progress requires that knowledge accrued through the experiments of some members of society be freely gifted.
The concentration of economic power and benefits is however increasing, based on depriving others of access to knowledge. Legal monopoly is already very well advanced, with just 2,000 corporations owning 60 per cent of the patents simultaneously obtained at the world’s five leading patent offices.
In the digital arena, secrecy is also a prevailing form of knowledge privatisation. Only 15 per cent of artificial-intelligence papers disclose the code involved. Google’s DeepMind is among those organisations that usually do not.
Additional mechanisms
Three additional mechanisms escalate global intellectual monopolisation. The first is predation in corporate-scientific networks. This is particularly evident in the pharmaceutical industry, where companies rely extensively on the work of scholars and use public funding for their research, yet they alone capture the profits from commercial exploitation.
A recent example is Remdesivir, used to treat Covid-19. This drug was patented and sold at an exhorbitant price by Gilead, even though entirely based on university research funded by the US National Institutes of Health. The NIH is the most frequent external funding source declared in Pfizer, Novartis and Roche scientific publications. Similarly, Google, Amazon and Microsoft co-authored between 78 and 87 per cent of their scientific publications until 2019, mostly with universities, but only shared ownership of between 0.1 and 0.3 per cent of their patents with other organisations.
A second, self-reinforcing intellectual-monopoly mechanism is related to the harvesting of data, where not only privacy concerns are at issue. In many industries, as the former Siemens chief executive Joe Kaeser said, manufacturing and engineering data are ‘the holy grail of innovation’. Since deep-learning algorithms learn and improve by themselves as they process more data, data harvesting results in continuous technical improvement.
Deep learning significantly automates discoveries and expands the types of problem that can be addressed through big-data analysis. Companies mastering this technology and exclusively owning original data sources expand their intellectual monopoly at an accelerating speed. This is true in many sectors, from finance with BlackRock’s platform Aladdin to retail with Walmart’s aggressive push towards proprietary data-analytics capacities. Nonetheless, technology giants increasingly occupy a leading role.
In 2015 Amazon, Microsoft, Google and Alibaba held in their public clouds around 4.9 per cent of the data stored worldwide but by 2020 this proportion had already reached 22.8 per cent. In their clouds, these companies offer deep-learning algorithms as a service. This means that, even without direct access to clients, algorithms can learn from third-party data, expanding the firms’ intellectual monopolies and allowing them to jump into other industries, from healthcare to transport.
A third phenomenon is related to the expansion of global value chains. The corollary of the unbundling of productive activities allowed by information and communication technologies is a dramatically increased circulation of information and a related sophistication of information systems, hand in hand with a concentration of the capabilities to govern networks. Lead firms’ planning capacity ranges from defining the dimensions of each production step taking place in subordinate companies to the setting of norms, standards and behavioural patterns. Moreover, the uneven distribution of uses of intangible assets along different nodes of the chains allows firms specialised in knowledge-intensive segments to capture most of the gains from scale economies.
Apple’s ‘fabless’ (outsourcing of fabrication) model and its masterful control over supply chains is a case point. The firm abandoned factories in Colorado Springs and Sacramento in 1996 and 2004 respectively, becoming the most renowned factory-less goods producer in the world. Most of its manufacturing is performed by firms in China and elsewhere in the global south, while Apple built ‘a closed ecosystem where it exerts control over nearly every piece of the supply chain, from design to retail store’. Critical in this panopticon surveying a highly dispersed manufacturing process is the monopoly over intellectual capabilities which allows Apple to capture the lion’s share of the value produced in the chain.
Falling short
Increasing awareness of the economic, social and political risks associated with the rising concentration of corporate power led to a recent antitrust push, first in the EU and the United Kingdom, followed by the United States and latterly China. Such moves however fall short of the challenges raised by intellectual monopoly capitalism, which go beyond the Big Tech companies and encompass much more than conventional market concentration.
What is at stake is a concentration of the ability to understand, co-ordinate and transform social and economic processes. Intellectual monopoly concerns new collective capabilities which should not be harnessed to the profit motive but rather be mobilised to achieve shared social, ecological and psychological development goals. This requires a new generation of resolute, innovative and co-ordinated policies, along at least two main dimensions.
First, following the ‘do no harm’ principle, extensive algorithmic accountability should prevail. Addressing responsibility for algorithmic decision-making should move beyond privacy issues and the biases leading to discriminatory and inequitable outcomes. Since control over algorithms makes it possible to ‘model, anticipate and pre-emptively affect possible behaviours’ and these capabilities are subjected to powerful monopoly forces, public authorities must prevent corporate uses of big data that encourage detrimental behaviours, such as compulsory consumption, carbon-intensive activities or online bullying. To that end, large-scale algorithmic apparatuses should be submitted to annual mandatory auditing, with publication of the relevant results.
Secondly, the resolution of crises and the attainment of socially and ecologically desirable goals must not be curtailed by intellectual monopoly. Patents should be automatically and generously waived where the free circulation of knowledge can contribute to alleviating social, health or ecological hardships.
Furthermore, intellectual monopolies set science and technology agendas, as evidenced by Big Pharma. This results in innovation rates and directions that privilege profit-making over solving social, ecological and health crises. Global institutional efforts are required to set new research agendas suported by public funds. But this is not nearly enough.
In the midst of the pandemic, Google made temporarily available its community mobility reports, which helped to assess the impact of mobility restriction on the spread of the disease. It is shocking that general-interest data such as these are not available on a permanent basis. Given the ability to process knowledge and behavioural data has become a powerful governing tool, algorithms should be open-source and data of general interest publicly available in anonymised form. Only thus can relevant big-data arrangements be deployed to serve public policies and prevent the predation of value across the economic landscape.
The Chinese state is already moving in this direction in the financial sector. As part of the implementation of its social-credit system, the central bank called data collected by internet platforms a ‘public good’, which should be disclosed and regulated more closely. Aversion to the lack of democracy and pervasive state surveillance in China is no excuse for letting crucial resources for the co-ordination of social life end up as a private monopoly. Creating a digital commons comprising data, algorithms and digital infrastructure could tackle both surveillance and data-driven intellectual monopolies. It would be a potential avenue for a socialisation empowering public agencies and private socio-economic actors alike.
Moving in these directions would imply a U-turn from the proprietary ideology of the previous fin de siècle. But it would just be a fair return to society. After all, Big Tech data scientists acknowledge that ‘the algorithms aren’t magic; they simply share with you what other people have already discovered’.