Leverage half-life turns enforcement from a snapshot into a race. A dependency that constrains an actor today — reliance on outside compute, energy, capital, or talent — can be routed around tomorrow as the actor integrates those layers into itself, and rising AI capability is a general engine for dissolving such bottlenecks. So the useful question is not “do we hold a chokepoint?” but “how fast is this one becoming routable?” The real test of a transition follows: is the ecology building durable constraint faster than the actor is dissolving the leverage that constraint depends on? A system that looks strong on paper may already be near impotence if the dependencies beneath it are disappearing. See constraining power without becoming it.
Leverage half-life
The rate at which a dependency that makes a constraint possible is becoming routable. Enforcement readiness is not whether the ecology holds a chokepoint today but how fast that chokepoint is decaying as an actor internalizes its own compute, energy, capital, and talent — so leverage is tracked as a trend, not a static checkbox.