For most of the past three decades, the direction of the global system was clear enough that it barely needed stating. More trade. More integration. More interdependence. Supply chains stretched across continents because that was where efficiency led them. Capital moved with relatively few constraints. The working assumption was that connection created stability, that shared economic interests were themselves a kind of order.
That assumption has not collapsed. But it is being quietly revised, and the revision is more significant than most of the commentary around it suggests.
The word that keeps appearing is deglobalisation. It is an understandable shorthand, but it misreads what is actually happening. Countries are not retreating from global trade. They are becoming more deliberate about where they depend and where they do not. That distinction sounds minor. It is not. It changes the underlying logic of how systems are designed, how risks are priced, and where investment flows.
The clearest evidence is in supply chains. For years, the dominant calculation was efficiency: if a component could be produced more cheaply elsewhere, it was. The complexity that resulted was manageable because the broader system was stable enough to absorb disruption. That calculation now includes a different set of variables. Resilience. Redundancy. Exposure to single points of failure. Production is being moved, duplicated, or regionalised not because it is cheaper, but because concentration has started to look like vulnerability.
Energy follows the same pattern. Access to reliable supply has shifted from an economic concern to a strategic one. Countries are investing in domestic capacity, securing long-term agreements, and reassessing dependencies that previously felt acceptable precisely because they were cheap. These relationships are becoming more political, which means they are also becoming less predictable and harder to price.
Technology is moving in the same direction, faster. Access to advanced chips, data infrastructure, and AI capability is increasingly tied to national strategy. What was recently a global market is being segmented through restrictions, incentives, and partnerships shaped by strategic interest rather than comparative advantage. Even capital, which remains global in scale, is more conditional in practice: risk is now assessed in geopolitical and regulatory terms alongside financial ones, and that changes where it goes and how quickly it moves.
A useful way to read this is not as fragmentation in the sense of collapse, but as a transition from tight integration to loose coupling. In a tightly integrated system, disruption in one part propagates quickly through the whole, which is efficient when things are stable and dangerous when they are not. In a loosely coupled system, connections still exist but are more contained. Dependencies are reduced where possible. Exposure is managed rather than assumed. The system trades some efficiency for a different kind of durability.
That transition is not seamless. Supply chains become more complex before they become more stable. Costs increase as redundancy is built in. Decisions that were once purely economic now carry political and strategic weight. And this is where the forces running through this series begin to intersect directly. Energy constraints make reliance on distant supply riskier. Inflation pressures increase the cost of disruption across every part of the chain. AI infrastructure concentrates in regions that can offer power and stability, which shapes where capability develops and who can access it. Each force reinforces the shift toward selectivity.
When systems are tightly integrated, coordination tends to emerge naturally from shared interest. As coupling loosens, coordination requires deliberate effort. Agreements need to be negotiated rather than assumed. Standards need to be actively aligned. Trust becomes both more important and more brittle. This is why geopolitical tension feels more persistent than episodic. It is not a series of individual disputes that will resolve when cooler heads prevail. It reflects a system in which alignment can no longer be taken as the default condition.
For organisations, this is often felt before it is fully understood. Supply chains behave less predictably. Costs fluctuate in ways that are harder to hedge. Regulatory environments diverge across regions in ways that multiply compliance complexity. The instinct is frequently to wait, to assume that conditions will stabilise and the previous model will reassert itself.
That may be the least useful response available.
The question is no longer simply about accessing global markets. It becomes about understanding which dependencies exist, which ones are load-bearing, and how exposure can be managed without retreating from the connections that still create value. The connections remain. But they now require effort that was previously unnecessary. And effort, in a system under pressure, is never free.
