You can sense that the internet is changing somewhere between the time a VPN connection drops in Shanghai and the moment a German user clicks past another cookie banner. It was once. It pretended to be, at least. The web marketed itself as a single, expansive commons for the majority of the late 1990s and early 2000s, where a teenager in Karachi and a librarian in Helsinki were, in a thin but genuine sense, neighbors. Now, and not by coincidence, that promise is eroding.
Governments have realized that the internet is power, something that the rest of us took years to acknowledge. Not symbolic power, but real, quantifiable, taxable, and observable power. Additionally, states typically want something to remain inside their borders once it gains power. The slow division of cyberspace into national enclosures is what people have begun to refer to as the “splinternet,” an ugly term for an even uglier concept. Ten years from now, we might discuss the open web in the same way that our grandparents discussed transatlantic telegrams—a brief, glittering period that came to an end in silence.
The most obvious example is China, which is the country that everyone aims for first. The Great Firewall has existed for so long that it almost seems like geography. The laptops in a Beijing café are open to WeChat, Baidu, and Douyin—a parallel ecosystem so comprehensive that most users don’t seem to notice what they can’t access. However, China is no longer the more fascinating tale. It’s because everyone else is beginning to take notes. Since enacting its sovereign internet law, Russia has been testing how much of the worldwide web it can shut off without going bankrupt. For more than ten years, Iran has been surreptitiously developing a national intranet. In a single move, India outlawed hundreds of Chinese apps and heavily favored domestic substitutes like Moj.
These actions are not motivated by ideology; democracies are also making similar moves, albeit with superior legal counsel. The goal of Europe’s GDPR was user protection, which is a legitimate goal rather than fragmentation. However, the actual result has been a subtle reorganization of how businesses produce goods. According to a friend who works for a midsize SaaS company in Lahore, they currently have three distinct data pipelines: one for the EU, one for the US, and one for the rest of the world. He said it in a casual manner, similar to how you would explain paying different utility bills. That’s the splinternet’s texture: a thousand tiny workarounds piling up instead of dramatic blackouts.
There’s a feeling that the early web developers either didn’t anticipate this or didn’t want to. With ICANN, IETF, engineers, and academics debating standards in fluorescent-lit hotel conference rooms, the previous governance model presumed that everyone had the same goals. For a while, they mostly did. Then the internet became too profitable, too dangerous, and too helpful to be left to the engineers. Governments have remained at the table since they arrived.
What we lose is the more difficult question. A split internet is more costly, brittle, and slower. Additionally, it makes the little magic of the old web seem unlikely, like sending a meme from São Paulo to Seoul without giving it any thought. It’s difficult not to question whether the borderless internet was ever stable in the first place or if it was always borrowed time as you watch this develop. Platforms and investors appear to be hedging both ways these days, preparing for fragmentation while openly lamenting it. The paradox is instructive.
Perhaps the global web was the exception. Given that the internet was created by humans, who frequently erect walls, it’s possible that the splinternet is simply the internet.
