Every major infrastructure revolution eventually produces its gatekeepers. Railroads did. Oil did. Telecommunications did. The digital economy is unlikely to be different.
What changes across time is not the underlying pattern, but the form power takes. In the late 19th century, America confronted the dominance of railroads because they controlled the routes through which goods, people and markets moved.
A farmer, trader or manufacturer could be economically free in theory, but dependent in practice on the terms set by a dominant rail corporation. Standard Oil followed by consolidating control over the systems through which oil was refined, transported and distributed.
Later came AT and T and telephone networks. Then came Microsoft’s dominance in operating systems, which reshaped the entire software industry. In each case, the dominance was not only over a product, it was over the road to the customer.
The lesson here is not that large companies are automatically bad. Many became dominant because they built superior products, invested early and served customers efficiently. The problem begins when control over infrastructure becomes control over market access itself. That is the question now confronting the digital economy.
A small number of firms now sit at the centre of cloud computing, search, app distribution, online advertising, social media, digital payments and more recently, artificial intelligence. Google, Amazon, Apple, Meta and Microsoft do not just sell digital services. They shape the conditions under which other businesses operate. Kenya lives inside this reality.
A retailer depends on advertising systems whose pricing and rules it does not control. A media business depends on search rankings and recommendation algorithms it does not control.
A bank that runs a significant share of its retail flows is dependent on infrastructure controlled elsewhere in the market. An insurer building underwriting tools on an external cloud provider depends on systems, service terms and outage risks it does not control.
Businesses remain independent in law. They become dependent in operation. The November 2025 anti-trust case brought by the US Federal Trade Commission (FTC) against Meta illustrates how difficult it becomes to regulate digital dominance once markets begin shifting beneath regulators.
The FTC argued that Meta maintained monopoly power through anticompetitive acquisitions, most notably Instagram and WhatsApp.
Meta’s defence was revealing. The company argued the market the regulators were trying to define no longer existed in the form they had imagined. Facebook and Instagram, it maintained, now compete with TikTok and YouTube in a broader battle for attention shaped by short-form video and algorithm-driven content.
The court ruled in Meta’s favour, accepting that digital markets had evolved beyond the narrower ‘personal social networking’ market the FTC had attempted to isolate.
The deeper issue is that digital infrastructure shifts form faster than regulation adapts. Railroads, pipelines and telephone networks were relatively easy to identify. Digital systems are more fluid. Social networks become video platforms.
Search companies become AI companies. E-commerce firms become cloud providers. App stores become payment gatekeepers. Cloud firms become foundations on which AI systems are built. By the time regulators define the market, the market may already have moved.
Kenya should therefore not attempt to copy American antitrust law mechanically. Our institutions, markets and development priorities are different.
That notwithstanding, the historical pattern matters. Every major infrastructure revolution produces gatekeepers before law and policy effectively adapt. The response should be strategic preparedness, not hostility to technology.
Regulators will need the analytical capacity to understand switching costs, platform dependency, cloud concentration and data control as structural economic questions. Procurement officers should understand that adopting a cloud, payments or AI system is rarely just a purchase decision. It is often the beginning of a long-term operational dependency.
General counsel will need to read technology contracts the way previous generations read infrastructure concessions or joint-venture agreements.
The next generation of monopoly questions may not involve a single railway line or telephone network. They may involve quieter systems: cloud infrastructure, payment rails, recommendation algorithms, app permissions and AI models. These are harder to see and govern, which is precisely why they accumulate power.
For Kenya, the real test will be whether the country can build the legal, commercial and institutional capacity to manage dependency before it hardens into vulnerability.
History does not warn us against innovation, it warns us against discovering too late that the infrastructure of opportunity has become the architecture of control.