Who Does the Internet Belong To?

On 30 April 1993, two directors at CERN signed an internal document. One was the Director of Research. The other was the Director of Administration. The document was addressed simply: "To whom it may concern."

It concerned almost everyone.

It released the software behind the World Wide Web into the public domain. CERN gave up its intellectual property rights and allowed anybody to use, duplicate, modify and redistribute the code. There was no charge, and no requirement to come back and ask what somebody else was allowed to build.

The Web had an owner. Its owner decided not to behave like one.

By the end of that year there were more than 500 known web servers. A tool originally proposed so that scientists could share information was becoming something nobody could completely contain.

More than thirty years later, around six billion people use the internet. It carries our conversations, businesses, photographs, relationships, education, entertainment and memories. Technically, no single person, government or company owns it.

That is the official answer. It is also an incomplete one, because most of us do not experience the internet as open code or a collection of independently connected networks. We experience it through places owned by somebody else.

The internet may not have an owner. It has landlords.

We stopped visiting and started moving in

The early Web was somewhere people visited. You moved between websites made by different people and organisations. A person could own a domain, build a page and decide what appeared on it. The experience was inconsistent, occasionally ugly and often confusing, but no single company could decide what the entire Web should show you next.

Over time, we stopped travelling as widely. The things we wanted gathered inside a smaller number of platforms. Our friends were there. The people we admired were there. So were the photographs, videos, conversations, shops and audiences. Convenience made the rooms fuller, and the fuller they became, the harder they were to leave.

On 9 April 2012, Facebook agreed to buy Instagram for approximately $1 billion. The company had built the software. Its users had filled it with reasons to return. Facebook was not buying the copyright to everybody's photographs. It was buying the platform that sat between those photographs and the people who wanted to see them.

That route was worth a billion dollars.

We create much of what makes platforms worth visiting: the photographs, jokes, arguments, tutorials, recommendations, dances, memories and conversations. Our attention attracts more attention, and attention attracts advertising. But creating the value does not give us control over the place where it lives. We can own a photograph and still not own the route by which anybody finds it.

A platform's most valuable product may be the people who believe they are only using it.

The first arrivals found land. The rest found platforms.

The internet did not reach humanity all at once. Its earliest users were researchers, government workers, universities and people with access to expensive computers. It moved into offices and wealthier households, particularly in high-income countries. Broadband widened the experience, and then smartphones brought a much larger, mobile-first population online across Asia, Latin America and Africa.

In 2010, 30% of the world was online. By 2020 it was 60%. By 2025, 74%.

People arriving at each stage did not meet the same internet. Earlier users were more likely to come through a computer. They learned browsers, websites, files, folders and search, and many experienced the internet as something they could build on. Later users were more likely to arrive through a phone and go straight into an app. For them, the internet could look like Facebook, WhatsApp, YouTube or TikTok rather than a wider system beyond them.

Facebook's Free Basics made that difference unusually visible. The service let people in participating countries use a limited set of websites without paying for mobile data, and it was presented as a way to connect people who might otherwise stay offline. But it was not the whole internet for free. It was a selection, and Facebook and its partners helped decide what was in it. For somebody who could not afford ordinary data, the company was not simply offering a way in. It was helping decide what the internet would contain once they arrived.

India effectively blocked Free Basics in 2016, when its telecoms regulator banned pricing that discriminated by online content. The argument was not only about price. It was about whether one company should be allowed to stand between a new user and the wider internet, then choose which parts were free enough to become familiar.

The first arrivals found somewhere to explore. The next found somewhere to join. Sometimes they were offered a version in which somebody else had already chosen the doors.

The internet looks universal from places where it nearly is

Around 2.2 billion people remain offline. That is not a small group waiting at the edge of modern life. It is more people than live in China.

In high-income countries, 94% of people use the internet. In low-income countries the figure is 23%, and across Africa it is 36%. The global average hides two entirely different realities: one where being offline has become unusual, and one where it is still completely normal.

The divide carries on inside countries too. Globally, 85% of people in urban areas are online, compared with 58% in rural areas. In low-income countries, only 14% of rural residents are connected. Women are still less likely to be online in many lower-income countries, and older people less than younger ones. Education, disability, income, infrastructure, language and conflict all shape who gets in.

More than 96% of the world now lives within reach of a mobile broadband network, yet only 74% is online. A signal can reach you without the internet being affordable, useful, safe or available in a language you understand.

Coverage tells us whether the door exists. It does not tell us who can afford to open it.

Online is doing too much work

Two connected people can have almost nothing in common. One has fibre broadband, unlimited data, several devices, a private place to work and the skills to build a business. Another shares a phone, rations prepaid data and relies on an unstable connection.

Both sit in the same column: online.

The statistic records that they got in. It tells us nothing about what they can reach, create or own once they are there.

A phone made publishing cheaper. It did not make creating equal. Content still needs data, storage, electricity, time, skills and usually somewhere suitable to work. Only 60% of internet users show even basic digital-content-creation skills, rural users lag behind, and education remains the strongest predictor of who has them. Lower-income users may be able to post. That is not the same as being able to produce consistently, build an audience or turn the work into a reliable income.

Even among people who do get paid, the rewards are concentrated. In 2025, the top 10% of creators received 62% of creator payments, up from 53% two years earlier.

We describe creators as independent while their relationship with their audience can depend entirely on a company they cannot telephone. The internet made publishing possible for almost anyone. It did not make creating equally possible, and it certainly did not make it equally profitable.

What race is the internet?

There is no reliable global breakdown of internet users by race. Race is not measured consistently around the world. Countries use different categories, many collect no racial data at all, and location cannot safely be converted into race.

That does not make race irrelevant to the internet. It means we have to ask better questions. Who uses it? Who built its largest platforms, and who owns them? Whose language became standard? Whose history is searchable? Whose culture moves across it, and who profits when it does?

Asia-Pacific holds more than half the world's population, and 77% of the region is now online. By number, the largest share of the internet's population is therefore likely to be Asian. Yet English is used by almost half of websites whose language can be identified, and UNESCO reports that just fourteen languages are present in more than 91% of web content.

The average internet user is unlikely to be white. The internet can still feel as though Western culture is its default. It may be Asian by population, American and Chinese in much of its corporate power, English in much of its searchable record, and heavily shaped by Black culture without Black people holding anything like an equivalent share of its ownership or profit.

In 2021, Black TikTok creators made that imbalance visible by refusing to choreograph a dance for Megan Thee Stallion's new song. The strike was a protest against a familiar pattern. Black creators originated dances that travelled around the world, while larger, often white creators got more of the attention, opportunities and credit for performing them. The absence became the evidence. When the people who usually supplied the movement stopped, everybody noticed what was missing.

Culture can travel much further than the name of the person who made it. That tells us more than any racial pie chart could. Who appears on the internet is not the same as who controls it, and who influences it is not necessarily who benefits.

Visibility is not ownership.

A phone can become almost anything

None of this should shrink the internet into an argument about harmful platforms or unequal access. It has widened human possibility in ways that are hard to overstate.

A phone can become a camera, classroom, shop, studio, map, library, workplace or piece of evidence. Families separated by migration can stay present in each other's lives. People can learn outside formal institutions, find communities they could never find locally and build businesses without first being chosen by a traditional gatekeeper. Someone can film what is happening in front of them and make the world see it. Artists find audiences without galleries, and musicians release work without labels. People with rare illnesses, marginalised identities or unusual interests find others who understand experiences that once felt entirely private. For many disabled people, the internet has opened new access to work, education, relationships and culture.

It lets ordinary people publish. That remains extraordinary.

The old gatekeepers did not disappear, though. Some were replaced by systems that are larger, less visible and harder to question. A newspaper editor could reject your work. An algorithm can make it almost disappear without ever telling you a decision was made.

Access is not ownership

My Instagram account was hacked when it had reached 57,000 followers.

It took me two months to get over it and start again. There was no dedicated help, and every message took weeks to get a reply. The follower count was not the whole loss. I had curated something that was special to me. It held photographs, work, conversations and a timeline of parts of my life. I had decided what belonged there and added to it over years.

Then it was gone.

The photographs were mine. The work was mine. The years it represented were mine. The place where I had put it all together was not.

Instagram let my work reach people who might never have seen it otherwise, and it helped me build an audience and a creative identity. It also kept the power to separate me from everything I had built there.

People often say that anything you put online lasts forever. What embarrasses you might. What matters to you can disappear by Tuesday. A platform can keep your data while removing your access to it, and keep holding traces of your life without giving you a reliable way to reach it.

We build there as though what we make is ours. The platform treats it as content on an account it can control.

Ownership can be designed differently

Wikipedia is also built mostly from work supplied by its users. Nearly 250,000 volunteers write, edit, fact-check, argue over sources and moderate its content. The Wikimedia Foundation hosts the infrastructure, but describes itself as a non-profit responsible for protecting Wikipedia rather than deciding what its articles say. There are no adverts between the paragraphs. Most of its money comes from readers. Its content is free to read and licensed so it can be reused.

It is not perfect. Its contributors do not represent humanity evenly, its articles inherit the gaps in available sources, and its internal decision-making can be complicated enough to need its own Wikipedia page. But its existence proves that scale does not automatically require the same bargain.

Wikipedia and Instagram both depend on people contributing things other people want to see. One turns those contributions into a shared public resource paid for by donations. The other turns them into attention sold to advertisers. The difference is not whether there are users. It is what the owner is for.

The internet remembers unevenly

The internet has become one of humanity's largest archives. Search engines use it to answer questions, researchers use it to understand behaviour, companies use it to predict what people want, and AI systems learn their patterns from enormous collections of digitised writing, images, audio and video.

What is online becomes easier to find, study and carry forward. What was never uploaded can start to look as though it never existed. A community can hold knowledge, language, history and sophisticated ways of understanding the world while leaving very little searchable material behind, and an archive cannot record what it never received.

That gives arriving late another cost. Earlier populations did not only have more time to build audiences and businesses. They had more time to turn their lives into data.

The next generation of technology may inherit the inequalities of the last. AI does not meet humanity evenly. It meets the parts of humanity that were recorded, digitised, made accessible and included in what it was given. The people missing from the internet can still be affected by decisions made with it.

They can be absent from the data and present in the consequences.

Nobody owns it. Somebody still decides.

No one owns the internet as a complete system. Different people own different layers of it. Telecoms companies own the networks. Technology companies own the servers and platforms. Governments set the laws and, in some places, control access. Shareholders own parts of the companies deciding what billions of people see. Users create much of the culture that makes all of it valuable, and our position is usually the least secure.

We can build an audience without being able to move it. We can create an archive without being able to recover it. We can supply the culture without owning the machinery that carries it.

That is why access cannot be the final measure of digital equality. The question is not simply whether somebody can connect. It is whether they can take part in their own language, understand the system, create rather than only consume, benefit from the value they produce and keep what they build.

In 1993, CERN gave up ownership so that other people could build without asking permission. More than thirty years later, billions of us build inside places where that permission can be withdrawn.

The internet is built by billions of us. The power to decide what survives belongs to far fewer.