There was a founding myth of the internet: that it would decentralize power, flatten hierarchies and make gatekeepers obsolete. The myth was powerful, and it is largely false.
What the internet actually did was create a new set of gatekeepers — smaller in number, less visible and more powerful than the ones it displaced. The critical layers of the network have concentrated into a handful of companies, and the concentration happened quietly, in infrastructure most users never see.
The myth of the flat network
The internet was designed to route around damage and to have no single point of control.
That design made the network resilient — and it also hid the commercial layer that grew on top of it. The underlying design distributes traffic; the commercial layer concentrates it. The two facts coexist: a decentralized technical network supporting an increasingly concentrated economic one.
This is the first thing to understand about network power: the architecture and the market tell opposite stories.
The layers nobody sees
The concentration is most extreme in the layers users never think about.
Domain names are managed through a hierarchy with a small number of critical registries and registrars. The domain name system itself — the lookup that turns a name into an address — runs on a small set of root servers and a larger but still limited set of authoritative services. Cables, data centers and content delivery networks are held by a handful of firms. The plumbing of the internet is oligopoly-shaped.
A failure or a policy decision at any of these layers can reshape what everyone on the network can do — without anyone asking the users.
The application chokepoints
Above the plumbing, the application layer has concentrated even more dramatically.
A small number of platforms control the way most people find information, connect with each other, discover content and conduct commerce. These platforms are the effective gateways to the public square. Their algorithms decide what is seen, their policies decide what is allowed, and their terms decide who can participate. They are infrastructure in everything but name — and they are treated as private companies with private rights.
The question of what obligations come with this kind of power is the central governance question of the digital age, and it remains unresolved.
The switching power
The most consequential form of network power is the power to switch — to redirect, restrict or reorder.
When a platform changes its algorithm, it silently redistributes attention across the economy. When a gatekeeper decides what can be listed, sold or promoted, it sets the terms of commerce. When an infrastructure provider decides which traffic to prioritize, it shapes the performance of competitors. These switching decisions are made daily, and their cumulative effect is enormous.
The people making these decisions are accountable to shareholders, not to the users and businesses affected by them.
The resilience risk
Concentration is not only a political problem; it is a technical risk.
A network with many points of concentration has many single points of failure — technical, commercial and political. A failure at one provider can take down vast swaths of the network. A decision at one company can reshape markets overnight. The resilience that the internet’s architecture promised has been undermined by the concentration of its economy.
The outages and incidents of recent years have repeatedly demonstrated this: the internet goes down in pieces, and the pieces are owned by a few.
The governance gap
The regulation of network power lags far behind the concentration itself.
Antitrust law was designed for an industrial economy, and its tools fit the digital one poorly. The rules that govern public utilities were designed for physical infrastructure. The platforms that function as public squares are treated as private property. The result is a governance gap: the institutions that exercise the most power over modern life are the least accountable to the public.
Closing the gap is not a technical problem; it is a political one, and it is being fought out in courts, legislatures and boardrooms around the world.
The honest assessment
The internet did not decentralize power. It re-centralized it — in new hands, at new layers, with new mechanisms.
That is not an argument that the internet has failed; it has transformed the world in ways that would have seemed impossible a generation ago. But it is an argument for looking clearly at where the power actually sits, and for treating the question of how that power is held accountable as the urgent question it is.
The network that was supposed to give everyone a voice has become, in its critical layers, a system run by a few. The concentration is not an accident of technology; it is the product of economics and regulation. And it is reversible — but only by people willing to look at the plumbing, and to ask who holds it.