Big Tech: The Companies That Own the Starting Point
A handful of companies now sit between most people and most of what they do online: searching, buying, messaging, watching, working. They are grouped under a vague label — big tech — that mostly describes their size rather than what they have in common.
What they actually have in common is more specific and more interesting. None of them owns most of what it sells.
The largest retailer by reach holds a modest fraction of the goods sold through it. The dominant video service produces a small share of what people watch on it. The search company writes almost none of the pages it returns. The app stores build very few of the apps.
What they own is the place people start. Understanding why that turned out to be the most valuable position in the economy explains almost everything else about them — including why they are hard to compete with, hard to regulate, and unusually nervous about things that look small.
Owning the starting point
Before the internet, the scarce thing in most industries was distribution — printing presses, broadcast licences, shelf space, delivery fleets. Owning distribution meant suppliers had to come through you, and you could set the terms.
The internet made distribution nearly free. When anyone can publish or list a product, distribution stops being the bottleneck, and the scarce resource moves to whatever is still limited: people's attention and their habits.
Whoever becomes the default starting point captures that scarcity. And once they have, suppliers have no real choice about whether to be there, because that is where the customers already are. This is aggregation theory, and it explains the shape of these businesses better than their size does.
Two things follow.
The businesses are asset-light relative to their power. Adding a million more listings, videos, or drivers costs the platform almost nothing, because it does not own them. Growth is not constrained by capital in the way a factory or a hotel chain is.
Suppliers get commoditised. Not through malice — through structure. When customers are loyal to the interface rather than to any individual seller, sellers become interchangeable and compete on price and ranking. The commission charged is really rent on access to aggregated demand.
The moats holding this in place are the ordinary ones: network effects where each user makes the service better for others, switching costs where years of data and habit accumulate, and scale economics where fixed costs spread across enormous user bases.
Why regulation keeps struggling
Competition law in most countries was built around a specific signature of monopoly: a dominant firm raises prices above what competition would allow, and consumers pay more. That test is measurable and has worked for a century on railways, oil, and telecoms.
It fits these companies badly. Many of their consumer services are free. Prices to users have not risen; in several cases the services have improved. On the classic test, there is no harm to find.
The power is real, but it shows up somewhere the old tests do not look: in the terms offered to suppliers, in the ability to favour one's own products in one's own ranking, in control over which businesses can reach customers at all, and in the accumulation of data that makes the position harder to challenge.
This is why cases against large platforms run for years and often end inconclusively. It is not usually incompetence. It is a genuine mismatch between the harm being alleged and the framework available to prove it.
Regulators have responded by writing new rules aimed at conduct directly — requiring interoperability, restricting self-preferencing, mandating that users can leave with their data. Whether these work is still an open question, and it is worth being honest that the record so far is mixed: some interventions have visibly changed behaviour, others have produced compliance theatre.
There is also a real tension that gets flattened in public debate. Some of what makes these services good — integration, defaults that work, the fact that everything talks to everything — is the same thing that makes them hard to compete with. Rules that break the integration to help competitors sometimes produce a worse product. That trade-off is genuine, and both "regulation will ruin it" and "regulation is obviously overdue" tend to skip past it.
What has actually displaced platforms before
The historical record is fairly clear, and it is not encouraging for anyone hoping to beat these companies at their own game.
Direct competition rarely works. Building a better version of a large platform's product and spending heavily to promote it has been tried repeatedly and mostly failed, because a better product with fewer users can genuinely be worse to use.
Platform shifts do work. Every large incumbent that has been displaced was displaced when the place people start changed — when the dominant interface moved and habits had to form again. In that window, incumbency partly resets, because the new default is briefly unclaimed. This is exactly the Innovator's Dilemma applied at the level of an entire computing platform.
That is why these companies behave the way they do around new interfaces. Buying small companies that look like nothing, investing heavily in categories with no revenue, reacting sharply to products that seem harmless — from outside it looks like paranoia or empire-building. From inside it is a rational response to the one thing that has historically ended companies in this position.
The open question now is whether the current shift in how people find and do things online represents one of those windows. If it does, the position of the current incumbents is less secure than their size suggests. If it does not — if it turns out to be a feature the existing platforms absorb — then the aggregation they already hold gets stronger, because they own the distribution any new capability has to travel through.
Both outcomes are consistent with what we can see today, and anyone certain which one is coming is guessing.
