Download
Details
The dataset includes both raw values sourced directly from original records and calculated values derived from these sources. Raw values include firearm manufacturing, imports, and exports, while calculated values, including firearms produced for the domestic market and cumulative totals, are based on formulas outlined in the data dictionary above.
Firearms produced for the domestic market account for those manufactured in the U.S. — less those exported to other countries — and firearms imported for sale in the U.S.
Some categories represent totals, while others are subsets. For example, the handguns variables include all handguns (both pistols and revolvers), so pistols and revolvers should not be summed with handguns to derive a total. This breakdown is only available from 1986 onward. Similarly, long_guns variables are the sum of rifles and shotguns. That means that adding long_guns to rifles and shotguns would result in a double count.
Citation
If you make use of this data, please include one of the following citations, preserving the link back to the Gun Violence Data Hub:
- In news copy — … according to data provided by the Gun Violence Data Hub.
- In news graphics — Source: Gun Violence Data Hub
- In research — The Trace Gun Violence Data Hub. (2025). Firearm Production
Limitations
Production vs. sales
Production figures reflect shipments to dealers rather than actual sales to end users. See our gun sales estimates for data on the latter.
Cumulative data for early years
ATF data from 1899 to 1945 is provided as an accumulated total rather than as annual figures because annual breakdowns are not available.
Police and military firearms
Military firearms are not included, but guns produced for federal, state, and local law enforcement are. Though we’re unable to differentiate between law enforcement and civilian guns, nearly all of the firearms accounted for in this dataset were likely purchased by civilians. The Small Arms Survey, a Switzerland-based outfit that publishes periodic reports on the global gun stock, estimated in 2018 that local, state, and federal police forces in the United States were armed with just over 1 million firearms, which would account for less than 1 percent of our guns-in-circulation estimate.
Reporting
Not all manufacturers consistently report their production numbers to the ATF; for example, about 30 percent of active licensed gunmakers did not file reports between 2016 and 2020, possibly leading to underestimation of total production. However, these are mostly small manufacturers that make up “a relatively small percentage” of total production, according to the ATF.
Delayed and revised data
There is often a long delay, at times more than two years, before a year’s manufacturing, import, and export data becomes available. The ATF also sometimes updates its data as new and late reports are received, meaning initial figures can be subject to revision. When those revisions occur, we will update this dataset page.
Exclusion of DIY firearms
3D-printed guns and guns assembled from kits are generally not included in the ATF data, although a 2022 rule may improve future reporting.
Category aggregation
Categories like handguns and long_guns are already aggregated (for instance, handguns include both pistols and revolvers), so summing their subcomponents could lead to double counting. For example, summing handguns, pistols, and revolvers would result in a double count because handguns already include both pistols and revolvers. Refer to the data dictionary.
Methodology
The goal of the dataset is to provide comprehensive annual production information and estimated accounting of the number of guns in circulation in the U.S.
This dataset compiles firearm manufacturing, import, and export data from 1899 to 2023 using multiple ATF reports on the U.S. firearms industry:
- Data for 1899–1985 come from the ATF’s 2000 report, “Commerce in Firearms in the United States”.
- Data for 1986–1999 come from the ATF’s 2021 “Firearms Commerce in the United States: Annual Statistical Update”.
- Data for 2000–2023 come from the ATF’s 2025 “National Firearms Commerce and Trafficking Assessment”.
A uniform 1 percent annual attrition rate is applied from 1946 onward, which may not capture variations because of changes in firearm durability, production surges (for instance, during the COVID-19 pandemic), or other factors. The rate is based on one researcher’s methodology and is intended to be used as an estimate. For more on that, read our story.