For most of the internet era, technology was implicitly global. The same chips, the same standards, the same platforms spanned borders. That world is ending, and what is replacing it is something I have come to think of as bifurcation: rival powers building separate, increasingly incompatible technology stacks — their own chips, their own software ecosystems, their own AI infrastructure, their own standards. This is not a temporary dispute. It is a structural realignment, and its costs will be paid by almost everyone.
The dynamic driving this is straightforward once you trace it. When one power restricts another's access to critical technology, the restricted side does not simply give up. It builds its own. It develops domestic alternatives to the chips it cannot buy, its own software to replace what it cannot use, its own ecosystem to reduce dependence on a rival it can no longer rely on. Each restriction accelerates the very independence it was meant to prevent, and the two sides drift further apart with every move.
The result is two parallel worlds of technology that increasingly do not interoperate. Systems built in one ecosystem do not necessarily work in the other. Standards diverge. The shared global infrastructure that everyone took for granted fragments into competing blocs, each self-contained, each suspicious of the other. We are watching a single connected technological world split into two, in slow motion, decision by decision.
I think the costs of this are widely underestimated, and they fall unevenly. The most obvious cost is efficiency: duplicating entire technology stacks is enormously wasteful compared to a single global system, and that waste shows up eventually in higher prices and slower progress for ordinary people. But the deeper cost is to everyone caught in the middle — the companies, the smaller countries, the users who now have to navigate two incompatible worlds, choose sides, or maintain costly compatibility with both. Bifurcation does not just divide the rivals. It taxes the bystanders.
There is also a real loss that is harder to quantify. A great deal of human progress has come from the free flow of ideas, talent, and technology across borders. When that flow is choked off in the name of security and competition, we lose some of the cross-pollination that drove the advances in the first place. Two separate worlds, each guarding its knowledge from the other, will likely produce less collectively than one connected world did — even if each side advances within its own walls. The fragmentation has a cost measured not just in money but in foregone discovery.
I do not want to be naive about why this is happening. There are genuine security concerns, and there are real reasons a country might not want its critical infrastructure dependent on a rival. I am not arguing that the impulse toward independence is irrational. I am arguing that we should be clear-eyed that the cumulative result of all these individually defensible decisions is a more divided, more expensive, and probably less innovative technological world — and that this outcome is being reached without anyone really choosing it as a goal.
What I find most sobering is the momentum. Bifurcation, once it starts, tends to be self-reinforcing. Each step toward separation makes the next step more necessary, as each side becomes less able to rely on the other and more invested in its own stack. Reversing it would require a level of trust that is currently in very short supply. The split may be far easier to start than to undo.
So the question I keep coming back to is whether the world ends up permanently divided into competing technological spheres, or whether some shared layer — some common standards, some interoperability, some channel of exchange — survives the fragmentation. The answer will shape not just the economics of technology but the texture of the entire century. Are we building two worlds that can still talk to each other, or two that increasingly cannot?
The above reflects my personal views only and is intended for informational and discussion purposes. It does not represent the position of any employer or organisation.