February 26, 2026

Economists use the term “market power” to describe how monopolies gain the ability to unilaterally influence or control the terms of trade within a market, particularly the ability to set prices. Over the last three decades, evidence has accumulated that “platform power” should be recognized as a new and particularly potent form of market power.

In fact, platform power might be the ultimate expression of market power – greater even than the historical monopolies of the Gilded Age. In recent years, five of the first public companies to pass the trillion-dollar mark have been powered by it. Throughout the Internet era, platform power has driven escalating social change and brought extensive cultural and political disruption in the process.

The precise economic mechanisms of platform power remain the subject of debate. However, there is consensus about its essential elements:

1. Platform power requires digital communications networks.
2. Its strength is based partly on network effects such as Metcalfe’s Law – the principle that linear increases in network scale can lead to exponential increases in value.
3. Platform businesses often rely on tactics such as standards lock-in to protect themselves from competition.
4. Platforms collect enormous amounts of detailed information about users.
5. Platforms exercise power by using algorithms and vast quantities of personal data to manage interactions between participants.

 

All successful platforms are built on the same basic foundation – each starts by providing an integrated collection of valuable digital services. For example, the earliest platform businesses developed operating systems and communications networks. As more mature versions of the platform business model continue to emerge, the most successful platforms tend to provide social media services and generate revenue from advertising and marketing services.

Does this make TikTok the ultimate expression of platform power? Or might there still be further growth and development ahead?

I believe significant further development is imminent. I say this because at least one major opportunity has yet to be fully exploited, and it’s a big one – government platforms.

Why this is the case becomes clear if you distill a simple definition of the term platform from the consensus described above: platforms are proprietary integrated collections of essential resources made widely available within a networked environment.

This simple definition describes all major digital platforms – from operating systems to video-sharing sites – and it highlights the fact that as more and more government services become available online, governments themselves are becoming digital platforms. It’s important to note that as this occurs, we face the risk of governments gaining, and potentially abusing, uniquely powerful new levers of control – the same mechanisms of control exploited by private sector platforms, and collectively referred to as platform power.

 

The Rise of Government Platforms

Estonia pioneered the concept of digital government in 1998 when they began work on their X-Roads secure data transfer layer, which provided a foundation for the digitization of commerce and government services. Today, every aspect of the Estonian government is accessible via the e-Estonia platform, which was built on X-Roads. Following suit, India’s comprehensive government platform, known as IndiaStack, debuted in 2010 and has now scaled to 1.4 billion users. We’re still in the earliest days of this transition, but many other governments are moving rapidly in this direction, including Brazil, Singapore, United Arab Emirates, South Korea, and Denmark.

The most significant government platform currently under development is China’s. While China has not implemented a fully-formed, comprehensive, digital government platform like Estonia or India, they’ve expressed a commitment to do so under the Digital China initiative.

Platform businesses have demonstrated an incredible power to re-shape the industries they operate in. How can we expect this model to transform government?

While many changes can be anticipated, one stands out. Just as Bitcoin liberated money from traditional geographic constraints, government platforms can also operate irrespective of national boundaries. For example, any citizen of any country can use the Internet to register a business in Estonia, gaining the advantages of EU membership. Furthermore, countries that successfully implement comprehensive digital government platforms can easily export their codebase for adaptation and use beyond their borders, as India seeks to do. Over the long term, this will benefit first-mover governments such as India and Estonia by extending a form of platform lock-in to those countries, corporations, and individuals that adopt their platforms.

This means the race for supremacy in government platforms is also a race for international political and economic supremacy. And, uncharacteristically, the U.S. government has not formally committed to winning this race.

However, there’s a good reason for this.

Platforms Threaten Democracy

Going back to the Privacy Act of 1974, it’s been the practice of federal agencies in the U.S. to keep individuals’ sensitive personal data siloed – secure even from other agencies – to prevent the government from leveraging its extensive collection of information about U.S. citizens against U.S. citizens. Although criticized as outdated and inefficient, this policy remains valid – essentially, it protects us from the government gaining and abusing platform power.

This is because the essence of platform power involves using algorithms and personal data to shape the terms of public debate. Chinese economists are well-acquainted with this notion and use the term huàyǔ quán, or “discourse power,” to describe it. This term emphasizes the ability of platforms to apply algorithms and extensive quantities of personal data to shape public conversations, to sway opinions, and ultimately to serve as a tool for political control.

In democracies, governments are categorically prohibited from exploiting this kind of influence over their own citizens – in the U.S., it would violate the First Amendment. To state it plainly, platform power is anti-democratic, and digital platforms now enable the most sophisticated propaganda tools ever devised. As such, digital platforms are fundamentally unsuitable as a basis for democratic government.

The consequences of platform power being used to manipulate public discourse are not purely theoretical. The rise of digital media platforms has fueled a prolonged decline in the world’s news and information ecosystem. As former FTC Chair Lina Khan explained in 2018:

“Dependence on Facebook and Google for traffic has led publishers to package news according to the dictates of the platforms’ algorithms … [a]n entity with the power to dictate the terms of distribution of news has the power to dictate the content of news.”¹

Most leading digital platforms exploit highly intrusive surveillance technologies and aggressively apply data analytics, including advanced AI. Combining these technologies with the power of government threatens to unlock an unprecedented level of digital surveillance and control – far more extreme than what we’ve already seen. It is the path to digital authoritarianism.

Why Network Architecture Matters

However, there are still alternative paths available to us.

Today, virtually all major social media platforms – particularly the ones dominating online news and information services – operate as centralized networks. This network architecture is necessary to support platform power because it concentrates control over the distribution of content and the collection of data, making it possible to algorithmically manage billions of interactions between users. Centralization is essential to platform power.

Conversely, decentralized networks distribute responsibility for filtering incoming content and directing outgoing content among individual participants. This allows decentralized platforms to provide useful services without handing control of the entire network over to a few powerful individuals at the center. In decentralized communications networks, individuals communicate directly with other individuals, unimpeded by opaque algorithmic influences outside of their control.

Agentic AI Enables Decentralized Platforms

Fortunately, emerging AI technologies combined with blockchain-based identity management greatly facilitate the development of advanced decentralized communications networks. In particular, agentic AI – which allows systems to act autonomously on behalf of users – enables fundamental structural solutions, such as Francis Fukuyama’s “middleware” strategy.²

Because agentic AI has the power to simplify user interactions with complex systems, and because it can be operated locally under the control of individual users, it has the power to facilitate the assembly of private decentralized communications networks capable of displacing the centralized social media platforms we’re familiar with today. Using locally-hosted AI agents, individuals and communities gain the ability to create custom-built communications networks unmediated by platform owners and their personal or corporate agendas.

Agentic AI makes this possible by bridging incompatible data formats, collecting and analyzing large amounts of incoming information, automatically filtering and amplifying content, enforcing privacy protections, administering distribution lists, and managing other technical requirements associated with maintaining secure distributed networks.

By this same principle, citizens empowered by agentic AI also gain the ability to assemble customized interfaces to publicly available government assets – such as informational web sites, official documents, digital archives, and access to services – and to customize them to meet their personal requirements and priorities. This can be done without requiring governments to update or re-factor any of their existing systems.

Several European experts on digital government recently published a paper entitled The Agentic State. It’s informed by the Estonian experience – lead authors include Ott Velsberg, Estonia’s Chief Data Officer and Luukas Ilves, Ukraine’s Minister of Digital Transformation and formerly Estonia’s Chief Information Officer. It builds on the early accomplishments of the Estonian model and recognizes the unique role agentic AI can play to help mitigate the risks of platform power and to re-distribute control from the center to the periphery.

Regarding the promise of agentic government platforms, they write:

“The economics are compelling. Instead of multimillion-dollar system overhauls that often fail, governments invest in modest API infrastructure and let competition drive interface innovation. … Critically, governments need not rebuild legacy systems; they only need to open them to agent-mediated access.”³

The Prospect of Decentralized Government Platforms

This vision of an Agentic State describes an entirely different kind of government platform – one that can be explicitly pro-democratic. The opportunity this presents, which has only become realistic within the last year or so, is to use agentic AI to assemble decentralized interfaces to centralized government resources without building in the oppressive characteristics of centralized digital platforms we’re currently familiar with. This can be done while preserving the many benefits of existing government platforms – such as unified identity services, secure data transmission and storage, improved access to services, and greater resilience to hostile external influences.

As a consequence, it’s possible for decentralized network interface applications – essentially AI agents – to provide individuals with facilitated access to government services. Such networks can rapidly “self-assemble” and evolve into coordinated government platforms as technology progresses. This can happen inexpensively and in a very short period of time. Such decentralized agentic government platforms would be capable of providing most of the benefits of centralized services, but without granting a small number of decision-makers the power to manipulate and control millions of users.

As the world’s democracies confront a rising tide of digital authoritarianism, we cannot be confident that current strategies, such as influencing global standards and implementing export controls for advanced technologies, will be sufficient to prevail. Instead, our top priority should be to compete against authoritarianism on our own terms – on the basis of our ability to empower and defend the democratic rights of individuals.

¹Lina Khan, “What Makes Tech Platforms So Powerful?,” Promarket (Apr. 5, 2018), available at https://www.promarket.org/2018/04/05/makes-tech-platforms-powerful/

²Fukuyama et al., “Report of the Working Group on Platform Scale,” Stanford Cyber Policy Center (Nov. 17, 2020), available at https://cyber.fsi.stanford.edu/publication/report-working-group-platform-scale

³Ilves et al., “The Agentic State – Rethinking Government for the Era of Agentic AI,” theagenticstate.org, (Oct. 2025) available at https://agenticstate.org/paper.html