The Splinternet and the Geopolitics of Data

The phrase splinternet captures a development that would have seemed unthinkable to the internet's founders: the single, global network is fracturing at its data layer into distinct regional spheres, each governed by its own rules about what data may reside where and flow to whom. This lesson steps back from the specific statutes to see the geopolitical forces driving the fragmentation, because a practitioner who understands the forces can anticipate the rules rather than merely react to them.

Start with the recognition that data has become an instrument of statecraft. Governments now regard the data generated within their borders the way earlier eras regarded strategic minerals or industrial capacity: as a national asset to be protected, developed, and, where advantageous, denied to rivals. This reframing has several roots. Security agencies see foreign-held data as a surveillance vulnerability and foreign access to domestic data as a threat. Economic ministries see data as the feedstock of the artificial-intelligence economy and want the value refined at home. Political leaders see control over information as control over a lever of power. When data carries all these meanings at once, it stops being a commercial commodity and becomes a matter of sovereignty, and sovereignty does not compromise easily.

The result is a world dividing into broad data spheres. Europe's sphere is organized around the individual and the rule of law, exporting its regulatory model outward so that its standards shape behavior far beyond its borders, a phenomenon sometimes called the Brussels effect. China's sphere is organized around the state, tightly controlled at the border, with data treated as an extension of national security and a domestic technology ecosystem cultivated behind the wall. The United States has historically championed free data flows, reflecting the commercial interests of its dominant technology firms, but has increasingly carved out exceptions on national-security grounds, restricting the flow of sensitive data to rival states and controlling the export of the advanced chips on which data processing depends. Around these poles, other nations position themselves, some borrowing Europe's template, some China's, some, like India, improvising a pragmatic blend that guards key sectors while keeping trade open.

The hardening of these spheres is reinforced by the entanglement of data policy with the broader technology conflict, particularly around semiconductors and artificial intelligence. Controls on the export of advanced chips and the tools to make them are, at bottom, controls on who can build large-scale compute, and therefore on where the most valuable data processing can happen. A country cut off from the most capable hardware cannot train the largest models on its data, which reshapes the incentives around where data and compute co-locate. Data-localization rules and hardware-export controls are two faces of the same coin: both are about controlling the geography of computation. This is the fullest sense of the course thesis. Export controls, whether on bytes crossing a border or on chips crossing an ocean, reshape the global compute map by determining what can be computed where.

A practitioner should resist two tempting but wrong conclusions. The first is that the fragmentation is a passing storm that will clear once trade tensions ease. It will not, because the motives behind it, security, economic advantage, political control, are structural and enduring, not tied to any single administration or dispute. The second wrong conclusion is that fragmentation means the end of global business. It does not. Global firms continue to operate across all the spheres; they simply do so through regionalized architectures that respect each sphere's rules. The task is not to escape the fragmentation but to build for it, treating the boundaries between spheres as fixed features of the landscape to be designed around, like mountains or coastlines.

There is also a competitive dimension worth naming. The fragmentation advantages incumbents with the capital to build parallel infrastructure in every sphere and disadvantages smaller entrants who cannot afford to. A startup that could once serve the world from a single cloud region now faces a thicket of jurisdictional requirements before it can operate globally, raising the barrier to entry. Paradoxically, laws often framed as checks on the power of large technology firms can entrench those firms, because only they can readily bear the cost of compliance across every sphere. The practitioner advising a smaller organization must weigh which spheres are worth entering at all, because the cost of full presence in every one may exceed the return.

Faraday's read is that the splinternet is the defining condition of the coming decade of digital infrastructure, and that the professionals who thrive will be those who stop treating it as a compliance nuisance and start treating it as the fundamental shape of the terrain. The map is fractured. The work is to navigate it deliberately.