Easier To Stop Than To Start

Scope

This note is not about disruption risk. Markets have always priced disruption — the probability that something stops working. What markets have not priced is restart complexity — the probability that something stopped cannot be restarted on any timeline that matters to the financing structure depending on it.

The asymmetry is new and the mechanism is specific. Modern systems have been optimized so aggressively for continuous operation that the slack required to restart from a cold state has been systematically removed. The efficiency gain that looked like a competitive advantage during normal conditions is a restart fragility tax during stress conditions. The system is far easier to stop than to start. And the financial structures built on top of these systems are priced as if restart is trivial.

It is not.

What Changed

Optimization destroys restart elasticity. That is the mechanism.

A system running with inventory buffers, redundant suppliers, excess labor capacity, and operational slack can absorb an interruption and resume without significant coordination overhead. The slack absorbs the shock. Restart is a matter of reopening the throttle.

A system running at peak efficiency has no slack to absorb anything. Every node is synchronized to every other node. Inventory is just-in-time. Labor is specialized and non-fungible. Software dependencies are layered and version-sensitive. Restart requires reassembling all of those synchronized dependencies simultaneously — not sequentially, not gradually, but all at once, because the system cannot operate in a partially restarted state.

That reassembly is not guaranteed. It is not fast. And it is not linear.

Power grids that trip under load require carefully sequenced restoration procedures that can take days. Starting too fast causes cascading failures worse than the original outage. Semiconductor fabrication facilities that go offline face recalibration cycles measured in weeks — the equipment is sensitive to environmental conditions that took months to stabilize and cannot simply be switched back on. Refineries idled for maintenance or disruption require heating, pressurization, and chemical stabilization procedures before any product moves. Shipping networks disrupted by route closures require not just ship repositioning but insurance reinstatement, port slot renegotiation, and crew recertification — all of which operate on independent timelines that do not synchronize automatically.

The Hormuz disruption is the clearest live demonstration. The strait has not simply paused and resumed. The physical restart problem — insurance markets that withdrew coverage, shipping operators that rerouted fleets, refinery buyers that switched suppliers, freight contracts that expired — has created a reassembly problem that is orders of magnitude more complex than the original disruption. Each party that adapted to the disruption created a new dependency structure that now has to be unwound to restart the original one. The restart is harder than the stop because the stop was passive and the restart requires active coordination across counterparties who now have competing interests in whether and how quickly the original system resumes.

The financial structures financing these systems have not updated for restart complexity. They are priced for disruption risk — the probability and duration of a stop — but not for restart fragility — the probability that a stopped system takes significantly longer to resume than historical precedent suggests.

The distinction matters enormously for credit. A disruption that lasts two weeks is a working capital event. A disruption that lasts four months because the restart is nonlinear is a solvency event for any financing structure with quarterly covenants, mark-to-market provisions, or refinancing windows in that window. The same physical disruption produces a fundamentally different financial outcome depending on whether restart is fast or slow.

The implicit assumption in most credit models is that restart is fast — that once the physical disruption resolves, the financial impact resolves proportionally. That assumption was reasonable when systems had slack. It is not reasonable when systems have been optimized to the point where restart requires a full reassembly of synchronized dependencies that cannot be forced on any timeline shorter than the slowest component allows.

The slowest component is always slower than the model assumes.

→ Restoration timelines after discrete disruption events — the gap between physical resolution and operational resumption is the direct measure of restart complexity; when restoration takes multiples of the disruption duration the restart fragility mechanism is operating

→ Credit covenants referencing "disruption duration" without a separate "restart timeline" clause — the absence of restart-specific language is the structural evidence that restart fragility is not priced

→ Insurance market behavior after route or facility closures — when coverage withdrawal becomes sticky and reinstatement requires re-rating, the restart has acquired a financial dependency that did not exist before the stop

→ Supplier switching during disruption and the cost of switching back — each adaptive move made during the stop creates a new dependency that must be unwound during restart; the unwind cost is the hidden restart tax

→ Labor recertification and retraining requirements after extended idling — specialized labor that cannot simply return to work after a long stop is a restart bottleneck that does not appear in standard operational risk models

→ Software and systems version drift during extended outages — dependencies that were synchronized at the time of the stop may no longer be compatible at restart; the re-synchronization cost is a restart fragility tax specific to digital infrastructure

→ Insurance market reinstatement timing after disruptions — war risk insurance, cargo insurance, and operational coverage that withdraws during disruption does not automatically reinstate when physical conditions normalize; the insurance restart timeline is often the binding constraint on the operational restart timeline

→ Refinery and industrial restart disclosures — public companies operating complex industrial facilities increasingly disclose restart timelines and costs; acceleration in those timelines is the direct measure of optimization-induced restart fragility

→ Semiconductor equipment servicing concentration — a small number of companies service the equipment inside leading-edge fabrication facilities; any disruption to that servicing ecosystem has nonlinear restart implications for global chip supply

→ Grid restoration procedures and timeline data — NERC and regional grid operators publish post-event restoration analysis; the trend in restoration timelines relative to disruption severity is the cleanest measure of grid restart elasticity

→ Shipping route reinstatement after closures — the time between physical route reopening and normal freight volume restoration captures the reassembly complexity of the logistics restart; persistent gaps confirm the restart fragility mechanism

→ Debt covenant and refinancing window clustering near major industrial facilities — financing structures with tight covenants or short refinancing windows tied to facilities with high restart complexity are the most exposed to the asymmetry between disruption duration and restart duration

Markets price the probability of disruption. They price the duration of disruption based on historical precedent from systems with slack. They do not price the possibility that a disruption in a fully optimized system produces a restart timeline that is nonlinear relative to the disruption itself — because the slack required to absorb and recover from the disruption has been removed by the optimization that made the system efficient in the first place.

The restart problem is not a new physical phenomenon. Complex systems have always been harder to restart than to stop. What is new is the scale of the optimization that has removed slack from systems simultaneously across power grids, semiconductor fabrication, refining, shipping, and software infrastructure. Each individual system was optimized rationally. The aggregate effect is a global infrastructure layer with dramatically reduced restart elasticity at a moment when the probability of disruption is rising across all of those systems simultaneously.

The Hormuz disruption is not just an energy story. It is a restart complexity demonstration running in real time at the most critical chokepoint in global energy infrastructure. The physical conditions have partially normalized. The operational restart has not. Insurance markets, shipping operators, refinery buyers, and freight contractors all adapted to the disruption by building new dependency structures that now resist unwinding. The original system cannot restart because too many of its participants have already restarted into a different configuration.

That is the asymmetry. The stop was passive. The restart requires active coordination across counterparties with competing interests in the outcome. The financing structures that assumed the restart would be as fast as historical precedent suggested were priced for a world with slack.

That world was optimized away.

Hampson Strategies — Market Note · May 13, 2026

Not investment advice. Personal observations based on publicly available data.

© 2026 Andrew C. Hampson II / Hampson Strategies. All rights reserved.

Full archive: hscai.org/market-notes · Institutional engagement: hscai.org · 865-236-1026