Path Dependency and the Future of War
How today’s system choices set the rails for tomorrow’s conflicts
Path dependency means early decisions about technology, standards, and governance set self-reinforcing trajectories. Once institutions, infrastructure, and doctrine adapt, reversing or switching tracks becomes prohibitively costly. What looks like a tactical procurement choice today can harden into the strategic character of conflict for decades.
We are already watching this dynamic take shape.
This analysis grew out of two starting points: the U.S. investment in seven AI defence categories, and the systems China showcased in its September 3 parade.
The United States has concentrated investment in seven categories of AI-native defence firms: integrated platforms that serve as the nervous system of modern war, swarms that make numbers decisive, production accelerators that collapse manufacturing timelines, infrastructure companies that govern data and trust, and generative giants embedding themselves into planning and decision-support. These choices are more than procurement; they establish rails where speed, orchestration, and machine-validated trust become the organizing principles of conflict.
China is laying down another set of rails: near-space platforms in the grey zone between air and orbit, cislunar logistics that extend maneuver into the Earth–Moon system, civil–military fusion that ensures every advance is dual-use, and a vast STEM pipeline that hardwires scale into its future options. These decisions favour industrial throughput, positional control, and the ability to write new rules in governance-light domains (I think this is an interesting pressure point).
Taken together, these trajectories point to a form of conflict defined by the systems themselves. The decisive factors will be the systems nations have locked into.
Wars will be fought on the cadence of updates and replenishment cycles - how quickly forces can refresh software and replace hardware will matter as much as battlefield maneuver.
They will be structured by credentialing systems - digital standards that determine which partners can connect and share data. These function as new forms of alliance architecture. Connection determines who has access to shared intelligence, targeting data, and command-and-control networks. The choice to align with one credential system over another sets long-term commitments: full members gain machine-speed integration, partial connectors operate at the margins, and non-participants risk exclusion altogether. In this sense, the systems themselves begin to redraw the lines of coalition warfare.
They will be shaped by simulation engines - warfighting models that over time start to guide doctrine, with the risk that militaries prepare for the world inside the model instead of the world outside it.
And they will be mediated by critical services - things like insurance, satellite launches, or equipment repair that are often controlled by private companies. If those services are denied or withdrawn, it can decide whether a mission succeeds or fails.
The risks are clear. Autonomous escalation loops could cause machines to push a crisis forward faster than leaders can step in to stop it. Misaligned optimization could drive entire campaigns toward narrow targets that don’t match strategic goals. Runaway countermeasure cycles could trap both sides in a costly spiral of action and reaction at machine speed. These are the foreseeable outcomes of the systems being built today.
This is why path dependency matters. Once institutions adapt to these systems, options narrow. Slowing down becomes unaffordable. Exiting a credential bloc becomes impossible. Overriding simulation norms becomes unthinkable. Strategy risks collapsing into maintenance of systems rather than purposeful choice.
The Next Step
If path dependency is real, then the question isn’t just what systems to build - it’s what kind of warfare those systems will create once they take root. That means slowing down long enough to understand the architectures, standards, and governance we are locking into before they harden. The danger isn’t only falling behind in the technology race; it’s winning on speed while losing the ability to choose our future.
