What Does It Take to Reindustrialize America?

What Does It Take to Reindustrialize America?

LEARN / REINDUSTRIALIZATION

What Does It Take to Reindustrialize America?

Nine things, and the country is not short of all of them. Capacity is set by whichever is scarcest, that input differs by industry and by region, and three of the nine have no federal measurement series at all.

Key points

The system

Nine inputs to productive capacity, and where each is measured Nine inputs feed industrial capacity. Bars show relative constraint; those marked with a hollow cap have no federal measurement series. Schematic; no quantities implied. SYSTEM CAPACITY Capital Demand Materials Infrastructure Workforce Permitting EquipmentUNMEASURED Supply depthUNMEASURED Energy BINDING Nine inputs to productive capacity, and where each is measured Nine inputs feed industrial capacity. Output is limited to the level of the scarcest. Three inputs have no federal measurement series. Schematic; no quantities implied. SYSTEM CAPACITY Capital Demand Materials Infrastructure Workforce Permitting Equipment Supply depth Energy BINDING FADED BARS: NO FEDERAL SERIES

Each of the nine feeds productive capacity, and the achievable level is set by whichever is scarcest. The arrangement above is schematic and implies no quantities. What it does show is the property that makes industrial capacity difficult to legislate: raising eight inputs while the ninth stays fixed produces no additional output, and spending distributed evenly across all nine spends most of the money on the ones that were not binding.

The nine, and what each entry found

Capital. Rarely the binding input in the United States, and the one most industrial policy is designed to supply. Private manufacturing construction spending ran at a $195.3 billion annual rate in January 2026, down from $223.1 billion in June 2025, and the Census series is not adjusted for price changes, so the real decline is larger than the nominal one.

Demand. A plant is built against expected orders. Where demand is uncertain over the asset life, the project is not evaluated and never appears as a constraint, because it was never proposed.

Materials. The United States is 100 percent net import reliant on yttrium, graphite, gallium, arsenic and tantalum. What binds is rarely extraction. Refining and separation concentrate in fewer facilities than mining does, and ore that must be shipped abroad to be made usable leaves the dependency where it was.

Infrastructure. Rail, road, water and port capacity serving a site. Adequate in most of the country and decisive in the places where it is not, which is a regional constraint that national figures cannot show.

Workforce. Manufacturing employment stood at 12.6 million in July 2026, roughly 35 percent below its June 1979 peak. The constraint is occupational rather than sectoral, and training duration is a lead time that capital does not compress.

Permitting. The Council on Environmental Quality reports that environmental impact statements account for roughly one percent of environmental analyses under NEPA, and has published both a 4.5 year average and a 2.4 year median for their completion. Most of the argument uses the timeline of the one percent.

Equipment. The machines that build factories are themselves manufactured, substantially abroad. This input binds the rate of expansion rather than the level of output, which is why it is invisible in assessments of what a country currently produces.

Supply-chain depth. A prime contractor knows its first tier by contract and its second by inspection. Below that the base is largely unmeasured, and concentration risk accumulates in exactly the tiers that no series reports.

Energy. NERC has raised its ten-year summer peak demand forecast from 55 GW in 2022 to 224 GW in the 2025 assessment. More than 2,060 GW of generation was awaiting grid connection at the end of 2025, with a median of over five years from request to operation and a historical completion rate near 13 percent. The Institute’s Energy & Power subject develops this input across eleven entries, from what the grid actually is to why installed capacity and deliverable energy come apart.

The binding input moves

A semiconductor fabrication plant in central Texas and a forging operation in western Pennsylvania are constrained by different things. The fab waits on power and on the equipment it cannot buy domestically. The forge waits on skilled operators and on the qualification cycle for a new part. Neither is short of capital, and a capital programme aimed at both helps neither.

The same movement occurs across time. An input that binds in one year stops binding when it is relieved, and the constraint relocates to whatever was next. Clearing an interconnection queue moves the constraint to transformer lead times. Clearing transformer lead times moves it to the electricians who install them. A policy that treats the current constraint as permanent will be aimed at the wrong thing by the time it takes effect.

Institute analysisCapacity policy should be organised around identifying and clearing the binding constraint for a specific industry in a specific place, rather than subsidising every input at once. The second approach is easier to legislate and easier to announce, and it spends most of its money on inputs that were not limiting anything.

What twelve entries found in common

Writing these entries produced one result that none of them was looking for. The constraints with the best federal measurement are consistently not the ones practitioners describe as binding, and the constraints that bind most often have no series at all.

Manufacturing employment, industrial production, capacity utilisation, construction spending and mineral import reliance are all published on a regular schedule by federal agencies, and all of them describe the aggregate. Meanwhile: load interconnection requests are excluded from the national queue dataset by design, so the wait a new factory faces for power appears nowhere. Supplier concentration below the second tier is not collected. Tooling lead time, qualification duration and startup yield are tracked inside firms and published by nobody.

Measurement noteThree of the nine inputs above have no national dataset, and a fourth, capability-level production, has none either. An assessment of American industrial capacity built entirely on published statistics is describing the portion of the system that was easiest to count, and the portion easiest to count is systematically the portion least likely to be a single point of failure.

Why this requires an institution

Closing that gap requires a standing function rather than a study, because the constraints move and a finding published once is stale within a year. The function has three parts that have to be performed together.

The first is measurement at the capability level. Whether a specific component can be produced at a required rate within a required time is a question no aggregate answers, and answering it requires working through particular supply chains rather than sampling the economy. The Institute for American Manufacturing & Technology builds that assessment program by program, which is what its work on industrial readiness for aerospace and defence is for.

The second is the operating layer. The people who know which input is binding are the ones running the plant: the operations manager who knows the forging house has one furnace, the controls engineer who knows the qualification took eleven months, the utility planner who knows the load study queue. That knowledge exists and is not written down anywhere a policymaker can reach. The Institute’s network exists to reach it, and the reference entries in this library are reviewed against it.

The third is publication under a standard. A finding that cannot be checked does not travel. Every figure in this library traces to a primary source named on the page, disagreements between official series are stated rather than resolved by preference, and the Institute’s own interpretations are marked as interpretations. That discipline is what allows a practitioner’s observation and a federal series to be weighed against each other in public rather than asserted past each other.

Institute analysisThe Institute for American Manufacturing & Technology exists because the binding constraints on American industrial capacity are largely unmeasured, and because measuring them requires both technical access to the operating layer and an editorial standard that keeps the result checkable. Neither alone is sufficient. A trade association has the access and not the independence. A university has the standard and not the access. An institution that holds both is the missing instrument, and this reference library is one of its outputs.

Common misconceptions

That reindustrialization is a spending question. Money relieves the inputs that money can buy. It does not shorten an apprenticeship, a qualification cycle, or an interconnection study, and those are frequently the ones binding.

That there is a national answer. The binding input differs by industry and by region. A policy calibrated to the national average is calibrated to a situation that exists nowhere.

That the aggregate statistics describe the problem. They describe the measured portion of it. Three of the nine inputs here have no federal series, and their absence from the data is not evidence of their absence from the system.

What the evidence says, and where it is contested

ContestedWhether industrial capacity should be pursued at all, and at what cost, is a genuine political question that this library does not settle. One position holds that a country trading freely with capable partners obtains goods more cheaply than it could make them, and that capacity for its own sake purchases insurance at a price rarely stated. A second holds that the scenarios worth insuring against are precisely those in which trade is interrupted, so the price of the insurance is the wrong test. The Institute for American Manufacturing & Technology takes the second view and marks it as a position. The measurement problem described above stands under either, because a country that decides to hold less capacity should still know how much it holds.

Related Institute research

Why Reindustrialization Requires More Electricity
The input currently binding in most regions.

Critical Supply-Chain Dependencies
Why concentration predicts disruption and import reliance does not.

SIOP for Aerospace and Defense
Capability-level industrial readiness in practice.

Powering the Buildout
The energy constraint quantified across three sources of new load.

Sources

Reference entry maintained by the Institute for American Manufacturing & Technology. Figures are drawn from primary sources and cited above. Where the Institute states a position rather than a finding, it is marked as such.

Private Professional Network

Institute for American Manufacturing & Technology

Join the IAMT Network. Connect with researchers, policymakers, and industry leaders working on American manufacturing, energy, and technology policy.

Apply for Membership →