Manufacturing Capacity Utilization Statistics (2026 Data)

By the editors at Reliable. Last updated October 20, 2026.

The short version: U.S. manufacturing operated at 75.9% of capacity in September 2026, according to the Federal Reserve's G.17 release of October 16, 2026. That is 2.3 percentage points below the 1972 to 2025 average of 78.2%, and the rate has been below that average in every month since May 2022.

Total industry (manufacturing, mining, and utilities) ran at 76.9%. Among the manufacturing industry groups, petroleum and coal products (88.3%), electrical equipment (86.9%), and machinery (84.3%) were running tightest. Furniture (66.5%), motor vehicles and parts (67.8%), and primary metals (68.0%) had the most unused capacity.

Abroad, EU industry reported 77.9% in July 2026 and China's manufacturers 73.3% in the third quarter, each measured by a different method. The Fed's annual revision, scheduled for November 24, 2026, may change the 2026 readings and the long-run averages quoted here.

At a glance

All U.S. figures are seasonally adjusted. The U.S., EU, and China figures come from three different measurement systems and are shown side by side for reference only. The differences are explained below.

What capacity utilization measures

Capacity utilization is the ratio of actual output to the maximum output a plant or industry could sustain. The Federal Reserve publishes the U.S. measure monthly in its Industrial Production and Capacity Utilization (G.17) release, which covers manufacturing, mining, and electric and gas utilities. All figures in the release are seasonally adjusted. The Fed's manufacturing aggregate is NAICS sectors 31 through 33 plus logging and traditional publishing, which were classified as manufacturing before the 2002 switch from SIC codes.

The Fed's capacity concept is "sustainable maximum output." In its G.17 Technical Q&As, the Board describes it as the highest level of output an industry can keep up on a realistic work schedule, allowing for normal downtime, and assuming the inputs needed to run the installed equipment are available. Capacity estimates that use a plant's current line speed or current shift pattern will differ from the Fed's. In the motor vehicle industry, where the Fed builds capacity plant by plant, it notes that industry estimates based on current configurations tend to run lower on capacity and higher on utilization than its own.

The main input to the Fed's capacity estimates is the Census Bureau's Quarterly Survey of Plant Capacity Utilization (QPC), conducted for the Federal Reserve Board and the Defense Logistics Agency. Each quarter, plants report the market value of actual production and an estimate of what they could have produced at full production capability. The survey instructions (form MQ-C2) tell respondents to count only equipment in place and ready to operate, excluding anything that would need extensive reconditioning; to assume normal downtime, maintenance, repair, and cleanup; to assume labor, materials, and utilities are fully available; to use the shifts, hours, and overtime that could be sustained under normal conditions and a realistic work schedule; and to hold the product mix typical of the quarter. The Fed folds QPC utilization rates into its monthly estimates as each quarter's survey arrives and into its capacity indexes at the annual revision. The Q1 and Q2 2026 tables were posted in June and September 2026.

Where U.S. manufacturing stands in 2026

Manufacturing utilization moved up 0.1 percentage point in September 2026 to 75.9%. Manufacturing output rose 0.3% in the month and 1.1% over the year, while manufacturing capacity grew 1.0% over the same twelve months, which is why the utilization rate has barely moved: output and capacity are growing at close to the same pace.

Source: Federal Reserve G.17, October 16, 2026 release, summary table and Table 12. Seasonally adjusted.

Beneath the flat headline, the composition shifted during the third quarter. Durable goods output rose at a 3.8% annual rate, carried in part by a 28.4% annualized jump in semiconductors and related electronic components, while nondurable output fell at a 1.6% annual rate. Motor vehicle and parts output dropped 3.0% in September alone.

Total industry utilization rose more, to 76.9% from 76.4% in August, because mining ran at 87.5% (2.3 points above its own long-run average of 85.2%) and utilities jumped 2.3 points to 73.8% on higher electric and natural gas output. Utilities remain 10.2 points below their 1972 to 2025 average of 84.0%.

Capacity utilization by manufacturing industry

The sector average hides a spread of more than 20 points between industry groups. The table lists the durable and nondurable manufacturing industry groups in the Fed's Table 7 for September 2026, plus the small non-NAICS group that completes the Fed's manufacturing total, with each group's 1972 to 2025 average and the gap between them. The share column is each industry's 2025 weight in total industrial capacity. The Fed's separate high-technology detail rows are discussed below the table.

Source: Federal Reserve G.17, Table 7, October 16, 2026. Seasonally adjusted. Gaps are Reliable's arithmetic from the published figures.

Six industries are running above their own long-run averages: nonmetallic mineral products (+8.0 points), machinery (+6.1), electrical equipment (+5.2), petroleum and coal products (+2.8), and, by small margins, aerospace (+0.4) and fabricated metals (+0.3). The two largest industries by capacity weight, chemicals (12.07% of industrial capacity) and food, beverage, and tobacco (11.93%), are both below average, at 72.3% and 77.3%.

The deepest shortfalls relative to history are in furniture (-10.9 points), paper (-9.2), primary metals (-9.1), printing (-8.5), and plastics and rubber (-7.0). Within the Fed's high-technology grouping, communications equipment stood at 58.4% against a 75.3% average, the lowest September reading among the manufacturing categories shown in the Fed's Table 7, while semiconductors ran at 76.5% and computers and peripheral equipment at 81.3%.

Historical highs and lows

The Fed's summary table carries a set of reference points for the manufacturing series, and Table 12 gives the monthly history. The two lows are worth keeping straight, because both are often cited as "the record."

Source: Federal Reserve G.17, October 16, 2026 release, summary table and Table 12.

The September 2026 rate is 13.3 points above the April 2020 reading and 12.5 points above the 2009 trough. It is 9.6 points below the 1988 to 1989 high.

Annual averages show how long the sector has run below its benchmark:

Source: Federal Reserve G.17, Table 12, October 16, 2026 release.

2018 was the last calendar year to average above the long-run figure. No calendar year in the Fed's published 2004 to 2025 table averaged 80% or higher; the highest in that span was 79.1% in 2007.

How the United States compares internationally

Three official series are available for 2026, and none is built the same way.

Sources: Federal Reserve G.17; European Commission DG ECFIN, July 2026 Business and Consumer Survey results; National Bureau of Statistics of China, Q3 2026 release.

The European Commission's Business and Consumer Survey results for July 2026 put EU industrial capacity utilization at 77.9%, down 0.1 point from April, and euro area utilization at 78.2%, down from 78.5%. The Commission's European Business Cycle Indicators report for the second quarter of 2026 (Technical Paper 089) puts the long-term average at 80.4% for both the EU and the euro area. The same July survey asked managers what was limiting production. Among EU industry managers, 33.9% cited insufficient demand as a factor limiting production, the lowest share since July 2023. Labor shortages were cited by 17.1%, material or equipment shortages by 12.6%, and financial constraints by 5.2%.

China's National Bureau of Statistics reported industrial capacity utilization of 73.1% for the third quarter of 2026, up 0.1 percentage point from the second quarter and down 1.5 points from a year earlier. Manufacturing utilization was 73.3%, down 1.5 points year over year. The release was published on October 19, 2026; the Bureau's English translation of its quarterly releases typically follows a few days later. The spread across China's listed manufacturing industries ran from 61.2% in non-metallic mineral products to 84.9% in chemical fibers. Computers, communication, and other electronic equipment stood at 80.2%, general-purpose machinery at 79.6%, and automobile manufacturing at 66.2%, down 7.1 points from a year earlier. The survey covers about 110,000 enterprises above a designated size and is not seasonally adjusted.

Two industries are worth lining up across the U.S. and Chinese releases. U.S. nonmetallic mineral products ran 8 points above their long-run average at 82.0%, while China's equivalent category was the lowest in its table at 61.2%. U.S. motor vehicles and parts at 67.8% and China's automobile manufacturing at 66.2% were both among the weaker readings in their respective tables. These are different measurement systems and different reference periods, so the levels should be read as directional.

What it means for maintenance and reliability teams

The Fed's capacity definition has a direct consequence for how maintenance shows up in the numbers. Capacity estimates already allow for normal maintenance and repair downtime, so a plant that takes its scheduled outages is still counted as capable of full production. Additional outages lower utilization when they reduce output during the measurement period while estimated capacity stays unchanged. Production recovered later in the period can offset some of that effect. For Hurricane Irma in September 2017, the Fed assumed producers recovered much of the lost output before the month ended, as its G.17 Technical Q&As describe. The Census QPC form asks plants operating below full production capability to mark the primary reasons, including insufficient supply of materials, lack of sufficient fuel or electric energy, and strikes or work stoppages.

Higher utilization indicates a smaller proportional gap between an industry's output and its estimated sustainable capacity. An industry at 88% (petroleum and coal products) or 87% (electrical equipment) has less aggregate room than one at 68% (primary metals) or 67% (furniture). For an individual plant, outage exposure depends on bottlenecks, inventories, available recovery time, and customer demand, none of which an industry average can show. That exposure has to be estimated from plant records. Our analysis of published downtime cost studies, Industrial Downtime Cost Benchmarks, covers how to build a plant-specific estimate.

At 75.9% utilization, measured capacity is about 32% above actual output (100 ÷ 75.9 = 1.32). That arithmetic describes the relationship between the Fed's two indexes with capacity held fixed. It does not establish how much additional production plants could deliver. Available capacity belongs to particular plants and processes, and the industry average cannot identify a plant's bottlenecks or its readiness to increase output.

Calculate your own

The plant-level formula is the one the Census Bureau asks respondents to use on the QPC:

Capacity utilization (%) = Actual production ÷ Full production capability × 100

Both terms are measured in market value of production for the quarter, at the same product mix. The Census instructions include a worked example in reverse: a plant with $1,200,000 of actual production that estimates it is running at 80% of capacity has a full production capability of $1,200,000 ÷ 0.80, or $1,500,000.

To align a plant calculation with the Census QPC concept:

1. Use equipment in place and ready to operate. Exclude facilities or equipment that would need extensive reconditioning before they could run.

2. Estimate the shifts, operating hours, and overtime that could be sustained under normal conditions and a realistic work schedule, assuming labor, materials, and utilities are fully available. Additional shifts count when they are sustainable; overtime hours do not count as an additional shift.

3. Estimate full-production output after allowing for normal downtime, maintenance, repairs, and cleanup. If full production would require added shifts or hours, allow for the downtime those hours bring with them. Do not deduct normal downtime a second time.

These assumptions align the plant calculation with the QPC definition. They do not make the result directly comparable with the Fed's published industry percentages. The Fed's Capacity Utilization Explanatory Notes describe additional estimation and aggregation steps, including adjustments that keep its rates consistent with the series' own history, and note that the rates the Fed reports tend to run higher than the rates in the Census surveys.

Capacity utilization and OEE answer different questions. OEE measures how well equipment performs during the time it is scheduled to run; capacity utilization measures how much of a plant's sustainable maximum output it is producing, including hours it chose not to schedule. A plant can post high OEE on a single shift while reporting low capacity utilization. Our guide How to Calculate OEE walks through the first measure.

Figures that circulate, checked against the source

"U.S. factories run at about 80% of capacity." The September 2026 rate was 75.9%. No calendar year in the Fed's published 2004 to 2025 history averaged 80% or more; the closest was 79.1% in 2007.

"Capacity utilization above 80% (or 82%, or 85%) signals inflation." The G.17 release does not specify an inflation threshold. What it publishes is the long-run average, currently 78.2% for manufacturing and 79.4% for total industry, and the gap between the current rate and that average. Historical research has estimated relationships between utilization and inflation. A 1996 Federal Reserve Bank of San Francisco Economic Letter, Capacity Utilization and Structural Change, treated 82% as the postwar average for manufacturing and the level at which inflation pressure was roughly neutral, while noting that the estimate was imprecise, with a 95% confidence interval running from about 78.5% to 83.5%. Estimates like that depend on the period and the model. They should be attributed to the particular study, and the Fed's statistical release should not be cited as their source.

"Manufacturing capacity utilization is at a record low." The lowest monthly reading in the Fed's Table 12 is 62.6% in April 2020. The pre-pandemic trough was 63.4% in June 2009. September 2026 at 75.9% is 13.3 points above the 2020 low.

"Capacity utilization is a direct measure of how hard equipment is running." It is an industry-level ratio of output to estimated sustainable capacity, built partly from survey responses. The Fed's own Technical Q&As note that for most industries it works from industry-level source data, with motor vehicles the exception, where capacity is built plant by plant from line speeds, shifts, and production days.

How we evaluated

Independent editorial analysis based on publicly available government data, peer-reviewed research, and the original pages behind widely quoted benchmarks.

Reliable Magazine does not sell scheduling software, staffing services, or consulting and has no commercial interest in routing readers toward any particular vendor. Reliable does not accept payment for rankings. Vendors may sponsor enhanced listings with additional detail, but editorial rankings are independent. Read our editorial policy.

Every rate in this guide traces to a named release: the Federal Reserve's G.17 of October 16, 2026 (summary table, Table 7, and Table 12), the Census Bureau's QPC program pages and form MQ-C2 instructions, the European Commission's July 2026 Business and Consumer Survey results and second-quarter 2026 European Business Cycle Indicators, the National Bureau of Statistics of China's third-quarter 2026 release, and a 1996 Federal Reserve Bank of San Francisco Economic Letter. Gaps and differences are our arithmetic from the published figures.

Honest limitations

The September 2026 figures are preliminary. The Fed revises each month's estimate in the five releases that follow; in the October release, August manufacturing utilization was revised from 75.7% to 75.8%.

A larger change is scheduled. The Fed's annual revision, scheduled for November 24, 2026, will move the base year of the indexes to 2022, incorporate Census benchmark data for 2023 and 2024, and fold in QPC data through the fourth quarter of 2025. Rebasing on its own does not change a utilization ratio, since output and capacity are rescaled together. New benchmark data and any changes in estimation methods can, and the Fed notes that method changes can affect an industry's index back to 1972. Individual 2026 readings and the long-run averages quoted here may change.

The Fed's "manufacturing" includes logging and publishing. Manufacturing on a strict NAICS 31-33 basis was also 75.9% in September 2026, with a long-run average of 78.1% instead of 78.2%.

The three international series measure different things. The Commission's figure is what managers report; the Fed's is an output index divided by a modeled capacity index; China's is a value ratio that is not seasonally adjusted and covers only enterprises above a size threshold. None of the three should be subtracted from another.

This guide quotes the Fed's industry rates, which are derived in part from the QPC. It does not reproduce the QPC's own industry-level full-production and national-emergency utilization rates for Q2 2026; readers who need those should download the quarterly tables from the Census page linked above. The Fed's rates are not the same as the survey's: its Explanatory Notes state that the rates it reports tend to be higher than the Census survey rates, because of adjustments that keep the Fed series consistent with its own history.

China's third-quarter release was published on October 19, 2026 in Chinese; this guide cites that page. The Bureau's English version had not been posted when this guide was prepared.

Industry aggregates say nothing about any individual plant. Within an industry at 68%, some plants are sold out and some are idle.

Related guides

• How to Calculate OEE: Formula, Examples, and Honest Limitations

• Industrial Downtime Cost Benchmarks: What Published Studies Actually Show

• The Real Cost of Unplanned Downtime in Manufacturing (2026 Data)

• Deferred Maintenance Backlog Statistics: Federal and Higher-Ed Data

• Manufacturing Employee Turnover Costs

Sources

• Board of Governors of the Federal Reserve System: Industrial Production and Capacity Utilization - G.17, October 16, 2026 (summary table, Table 7, and Table 12; the G.17 Technical Q&As, Capacity Utilization Explanatory Notes, and release calendar are on the same site)

• Federal Reserve Bank of St. Louis: Capacity Utilization: Manufacturing (SIC), CUMFNS (the Fed's broad manufacturing series, monthly from 1972)

• U.S. Census Bureau: 2026 Quarterly Survey of Plant Capacity Utilization (Q1 and Q2 2026 tables; the form MQ-C2 instructions in the program's technical documentation define full production capability and the worked example)

• European Commission, Directorate-General for Economic and Financial Affairs: Business and consumer survey results for July 2026 (EU and euro area capacity utilisation and limiting factors; the 80.4% long-term average is in European Business Cycle Indicators, 2nd Quarter 2026, Technical Paper 089, on the same site)

• National Bureau of Statistics of China: Industrial Capacity Utilization Rate in the Third Quarter of 2026 (October 19, 2026 release in Chinese, with industry detail and methodology notes)

• Federal Reserve Bank of San Francisco: Capacity Utilization and Structural Change (Economic Letter 1996-34, Joe Mattey, November 15, 1996)