The Splinternet Has Arrived

In March 2000, US President Bill Clinton likened China’s efforts at internet censorship to nailing Jell-O to a wall. “Good luck,” America’s leader joked in a White House speech. Beijing was then in the midst of constructing its Great Firewall. But Clinton, like many others, saw the internet as an unruly new frontier. The novel technology, he argued, would naturally push societies to open up. 

Years later, the Arab Spring was simulcast all over social media. Anyone with a smartphone — from anywhere — could intimately follow and engage with the region’s pro-democracy protestors in real-time as they tried to topple ossified illiberal regimes. 

“I believed that giving users such a simple way to navigate the internet would unlock creativity and collaboration on a global scale,” reflected Sir Tim Berners-Lee, creator of the World Wide Web, in an essay last September. “If you could put anything on it, then after a while, it would have everything on it.”

But today the internet is increasingly a realm where powerful actors exert control and malign groups exploit vulnerabilities. 

China’s rise as a technological power rests in part on Beijing having built much of the developing world’s digital infrastructure. This includes exporting its panopticon of state surveillance technologies. Brittle autocracies are reflexively throttling internet access for self-preservation. Organized hacking enterprises have grown cybercrime into the world’s third largest economy. The commercial spyware industry has exploded. Social media companies — helmed by billionaire owners — place users in algorithmically defined echo chambers and information silos. 

Renewed geopolitical hostility hasn’t helped either. Countries everywhere are racing to ditch their dependence on external providers of core services. Policymakers are thus barrelling forward on costly projects meant to advance digital sovereignty. Yet the concept remains ill-defined and technically daunting. 

In 2001, Clyde Wayne Crews, the then director of technology studies at the libertarian Cato Institute think tank, coined the term “splinternet” in a piece for Forbes magazine. He argued that hiving off global cyberspace into distinct competing networks would be more effective long-term. The world’s dizzying array of users, including businesses and institutions, could choose the network whose governance, protocols and infrastructure best suited their needs. Demand would shape supply.

A quarter century later, the splinternet is now increasingly a reality — just not for the reasons Crews hoped for. Instead, its manifestation is being driven by authoritarian retrenchment, surging protectionism and private tech firms morphing into weighty global actors. And it’s all poised to get worse. 

So are the risks and consequences. Mounting global challenges demand greater international coordination and understanding, not less.  

The Unraveling of the Digital Commons 

“Instead of one globally connected internet, we are seeing the rise of regionally governed digital spheres shaped by politics, security concerns and economic competition,” said Francesca Musiani, research director at the French National Center for Scientific Research, over email. She co-authored a 2022 study for the European Parliament forecasting how recently enshrined EU legislation — the Digital Services Act, Digital Markets Act and AI Act — may help address patterns of fragmentation. But also, how they might produce unintended consequences. 

“The fragmentation is not only technical but ideological,” Musiani warns. “Democracies and authoritarian states are developing different norms around speech moderation, surveillance, encryption and online identity.” She points to digital sovereignty as another key driver of divergence. “While these policies are often justified in terms of cybersecurity and privacy, they also create barriers to interoperability and reinforce geopolitical blocs.”

“There’s a mad scramble by governments to try to bring a globalized cyberspace into a territorial jurisdiction,” echoed Dr. Milton L. Mueller, a professor at Georgia Tech University and the co-founder and director of the Internet Governance Project, when asked for comment. “It can never work. But in the meantime, as states try it, they trigger destructive power struggles over information flows, trade in digital services and hardware, and cybersecurity.”

In its latest annual report on online liberties, US-based democracy watchdog Freedom House describes how these are being steadily eroded. It’s a consistent, ominous trend dating back more than a decade. Authoritarian leaders especially view micromanaging their digital ecosystems as key to regime stability. Gulf monarchies, for example, are still haunted by the Arab Spring uprisings that caught their neighbouring dictatorships off guard.

Egypt, Pakistan, Turkey and Venezuela have regressed the most in recent years, says Freedom House. Sudan, Myanmar and Vietnam rank among the most restrictive as well. Likewise for Russia, where citizens inhabit a highly controlled information environment — even if many people instinctively see through the propaganda . 

Yet China remains the most repressive overall. Beijing is harnessing new AI-powered tools to effectively rewrite China’s social contract. Abolishing privacy, citizens are told, will improve service delivery. Government policy also dictates domestic websites and internet platforms promote the hardline ideology of the Chinese Communist Party and personal thought of President Xi Jinping.

“Authoritarian regimes can fail at almost everything so long as they suppress all other political alternatives,” historian Stephen Kotkin has cautioned. 

Yet democracies aren’t immune to backsliding either. Indeed, many sit among the at least 85 countries that have triggered internet shutdowns as a means of social control — primarily to curb outbreaks of civil unrest. In India, the world’s largest democracy, the government has ordered internet blackouts at least 166 times. Each episode lasting nearly six days, on average.

Some governments are now going further, assembling their own digital fiefdoms by replacing the internet altogether.

Moscow earlier this year blocked WhatsApp and Telegram in a bid to push users toward Max, an unencrypted, state-backed domestic super app now preloaded on new smartphones across Russia. Rolling internet shutdowns have meanwhile become commonplace under the justification of wartime security. However, the Kremlin was hard at work insulating the country’s digital ecosystem long before launching its full-scale war in Ukraine.

In Iran, a brutal crackdown against pro-democracy protestors in January  provided similar cover to advance longstanding government efforts to sever the country from the global internet. Inserted in its place has been a tightly controlled intranet, dubbed the National Information Network. This domestic alternative consists of a small mix of rudimentary apps, government webpages and platforms for banking, health care and other basic services. 

“The result has been a near-total information blackout,” CBC reported in April. “The paucity of information means many Iranians didn’t learn about the scale of the January massacres until weeks later — in a country nearly the size of Western Europe, there is little way for news to spread by word of mouth.” 

Piecemeal internet access began being restored in Iran as of late May. At the same time, a senior regime official revealed that Tehran has imported Chinese technology to enhance its censorship capabilities going forward. 

AI and the Fragmentation of the Internet

The rapid evolution of artificial intelligence is also  coming into play. Nowhere is this more apparent than in the great power competition between China and the US. 

Nations throughout the Global South are getting locked into the Chinese tech stack. Beijing already offers a full package of IT hardware, software, training and financing to nations in Africa, Asia, the Middle East and Latin America via its Digital Silk Road initiative. Client states can now seamlessly integrate free, high-performance Chinese AI tools into local systems that are based on Chinese networks. Indeed, by late 2025 Chinese open-source models overtook American ones in global popularity. 

Washington and its Western allies are tracing their own paths as well.

The Trump administration’s AI Action Plan, released last July, takes a predictably ultranationalist approach. “America is the country that started the AI race,” Trump said in a speech the evening the plan was published. “And as president of the United States, I’m here today to declare that America is going to win it.” 

The strategy vows that existing government guardrails will be rolled back and “woke” safety measures will disappear. Federal funds will be withheld from states attempting to regulate AI on their own. Environmental laws that impede data centre construction are on the chopping block. Protectionist trade policies will be paired with efforts to champion American companies abroad. Vague details for mitigating national security risks arising from cutting-edge models only appear on the manifesto’s final two pages.

“Even longstanding allies and partners are framed as mere markets to be captured in the interests of global American supremacy,” Kat Duffy, a senior fellow at the Council on Foreign Relations, wrote alongside colleagues in an initial assessment of Trump’s AI strategy. “Ultimately, when it comes to global engagement, the plan sings a zero-sum tune.” 

The America First approach is thus stoking fears of economic coercion by present and future US administrations. And not only in Ottawa but also Canberra, Tokyo, Seoul and European capitals. The same is true for the previously obscure 2018 CLOUD Act. The extraterritorial legislation dictates that foreign servers owned or operated by American companies remain subject to US authority. 

A consortium of Europe’s top companies have responded by promoting the EuroStack initiative. The project is meant to ensure the continent’s digital independence by formulating a joint digital industrial policy plan that encompasses digital connectivity, cloud computing, AI and online platforms. 

Private tech companies have also become their own formidable actors, capable of unilaterally shaping the online environment to suit their own business needs. “Tech companies are adapting products to regional political demands, effectively creating separate internet experiences for different populations,” says Francesca Musiani. “Over time, this may produce parallel ecosystems with limited compatibility.”

And then there’s the emergence of AI web search summaries. Data shows these are slowing organic, free-roaming web traffic to a crawl by enabling “zero-click” behaviour — a term referring to when users don’t leave the initial search results page. Findings from the PEW Research Center indicate only 1 percent of users click on links cited within Google’s AI overviews. The use of large language models can be even more isolating. Analytics firm Tollbit has previously found 99.63% of users do not click on links to external sources of information provided within chatbot conversations.

The Costs of Fragmentation

This splintering of the global online landscape carries grave risks. 

“Climate research, public health monitoring, cybersecurity intelligence, and AI safety all rely on large-scale international data sharing,” points out Francesca Musiani.  “If countries restrict access to data or isolate their digital infrastructure, global institutions may struggle to build accurate models, coordinate responses, or verify compliance with international agreements.” Friction within the global economy will vastly increase, she predicts. Companies engaged in cross-border trade will more and more be forced to comply with incompatible privacy frameworks or navigate a mess of technical standards.

The result will be a negative, self-reinforcing cycle, anticipates Konstantinos Komaitis, a resident senior fellow with the Atlantic Council’s Democracy + Tech Initiative. “The shift from an open internet to a fragmented one represents a move from a global commons to a collection of digital islands,” he argued in a March essay. As countries try to carve out their own pockets of digital sovereignty, he writes, it risks accelerating the very instability that they fear. “Interoperability declines, costs rise, and cooperation becomes harder precisely when global coordination is most necessary.”

Komaitis proposes a range of solutions, all of which hinge on greater leadership from so-called middle powers. He argues a coalition of willing countries should commit to deepening internet interoperability as their default design of digital systems. Other nations would be welcome to join. Among other things, this grouping should preserve key aspects of open-source digital infrastructure. Take for example the Internet Archive and the Internet Engineering Task Force, a volunteer-based standards development organization. Much of this could be done through modular computing, Komaitis suggests. He also encourages shared audits and establishing common digital sandboxes for experimentation with AI, as well as data and platform governance. 

In other words: to save the open internet demands doubling-down — in apolitical ways — on its most fundamental strengths. The spin-off benefit: doing so may also give oxygen to some of the scant remaining forms of international cooperation. But it will require restraint on behalf of governments that are habitually partisan.

“We are seeing a massive breakdown of key institutions that provided what few successes we have in global governance following the Second World War,” said Professor Milton Mueller, from the Internet Governance Project. An important exception to this, he says, are elements crucial to web governance, such as the Domain Name System and regional internet registries.  One might also add the Internet Corporation for Assigned Names and Numbers, or ICANN. Founded in 1998, the California-based non-profit administers the internet’s “phone book” based on a multistakeholder model.  

These entities “have remained truly global and — not coincidentally — non-governmental in nature,” says Mueller. “Globalism lives there.” 

The original promise of the internet rested on the idea that a globally connected information system could expand access to knowledge and bolster communication and cooperation across borders. Increasingly, however, the online world is moving in the opposite direction: toward rival digital blocs shaped by geopolitical competition, platform power and AI-mediated systems of information.  One glaring risk is   a more disjointed and dysfunctional  internet. The  other  is a more fragmented global public sphere itself. Establishing shared facts, common standards and strengthening international coordination will become harder to sustain — precisely when global crises demand them most.