Production Got Cheap. Everything Else Didn't.

Scope

The last thirty years had one objective function: reduce the cost of producing things.

It worked. And the model that made it work assumed that everything else — maintenance, switching, verification, trust — would remain a background condition. Abundant. Frictionless. Effectively free.

That assumption is now the constraint.

This note is not about any single disruption. It is about the regime underneath the disruptions — the one the market is still reading as separate headlines.

What Changed

1. Switching cost is no longer negligible.

The AI infrastructure buildout is hitting a constraint that is not compute and not permitting. Large power transformers — the hardware that routes electricity from generation to load — are running delivery windows north of fifty-two weeks from a manufacturer base concentrated in single digits globally. The buildout was modeled assuming components would be available on demand because they always were. The switching cost from one supply architecture to another is now measured in years, not quarters. Capital authorization and physical delivery are not the same variable. Markets are pricing them as if they are.

2. Maintenance cost is now a primary transmission channel.

The Hormuz disruption is transmitting into industrial maintenance inputs — Group III base oils, synthetic lubricants, refinery feedstocks. Roughly forty-four percent of US Group III supply has been sidelined. This is not an energy price story. It is a maintenance-input story. The systems optimized for utilization never modeled the cost of keeping themselves running when the inputs for maintenance become scarce. That cost is now transmitting through fleet operators, logistics margins, and industrial uptime in ways that conventional energy-beta models do not capture.

3. Industrial surge capacity is not a function of budget authorization.

Defense spending is expanding faster than defense industrial capacity can absorb it. The constraint is physical — specialized chemicals, precision machining, thin supplier ecosystems built for peacetime procurement cycles, not sustained production demand. Industrial exhaustion precedes fiscal exhaustion. It does not appear in top-line revenue until well after it has already compressed throughput.

4. Trust has developed a cost structure.

The cost of producing information, credentials, and documentation is collapsing toward zero. Verification capacity is not scaling at the same rate. Organizations that modeled trust as a free input are discovering it carries a cost — in compliance infrastructure, in provenance systems, in authentication overhead. That cost is not yet in any consensus model for the businesses that will bear it.

The production economics are still real. The efficiency gains are not being reversed. What changed is that the system surrounding production — the maintenance layer, the switching layer, the verification layer — was never priced because it was never stressed. It is being stressed now, simultaneously, across domains that look unrelated from the surface.

The market is reading individual headlines. Transformer delays are an AI infrastructure story. Base oil tightening is an energy disruption story. Munitions throughput is a defense budget story. Verification costs are a technology story.

The constraint is the same in each case.

For thirty years, the binding constraint was the cost of producing. Capital, labor, and technology all moved toward minimizing it. The model worked because maintenance was cheap, switching was fast, and trust was free.

Those are not currently true simultaneously.

The regime that is assembling is not the opposite of efficiency. It is the tax on efficiency that was never modeled — the cost of maintaining systems that were optimized to produce, switching them when they fail, and verifying that what they produce is real.

Production got cheap. The constraint moved.

The market is still pricing production.

This note does not claim the production era is ending. It claims the unmodeled costs adjacent to production are now large enough to be the marginal constraint on returns — and that they are not yet in the models pricing the assets exposed to them.

The mechanism compounds if multiple maintenance, switching, and verification constraints activate simultaneously across the same capital structure. That is the scenario this note flags. It is not a forecast of when. It is a description of what is already accumulating.

Andrew C. Hampson II is the founder and principal of Hampson Strategies, an independent macro research and strategic intelligence practice. This note is for informational purposes only and does not constitute investment advice. hscai.org | @drampson11

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