10.3 million guns have been sold across the U.S. in 2025 through September 30, according to our tracker.
That includes 6.4 million handguns and 3.9 million long guns. The numbers are estimates calculated by The Trace using FBI figures. Year-to-date sales are down about 4 percent from the 2024 figure, and down considerably from the gun-buying boom during the pandemic.
Drilling down to the state level, we find some interesting trends. Sales are up by 8 percent or more in Rhode Island, Washington, Washington, D.C., and Nevada. Colorado has seen a rise of 4 percent.
Several of those places saw gun legislation enacted in 2025, among them:
— Rhode Island, which banned sales of new military-style semi-automatic rifles, shotguns, and handguns
— Washington state, where a permit will be required to purchase guns
— Colorado, where over a dozen bills relating to guns passed and were signed into law
Nevada’s Legislature advanced three bills relating to gun violence, though all of them were ultimately vetoed by Governor Joe Lombardo, a Republican.
Most U.S. states have seen dips in gun buying compared to 2024; in Oregon, Michigan, New Hampshire, and New Mexico that percentage point decline is in double digits.
Americans bought an estimated 1.2 million guns in December 2025, according to an analysis of FBI data.
This seasonally adjusted figure includes about 720,000 handguns and 480,000 long guns (rifles and shotguns).
The Trace’s firearm sales estimates are derived from data from the National Instant Criminal Background Check System (NICS), run by the FBI. NICS launched in November 1998; these estimates begin in January 2000.
Our analysis draws on the work of economist Jurgen Brauer of Small Arms Analytics, as well as BuzzFeed News and The New York Times.
Background checks, as the FBI notes in its published data, do not correspond one-to-one with sales, but in most cases, background checks for firearms transfers are conducted by licensed firearms dealers at the point of sale. We use these transfer checks to produce estimates of gun sales, taking into account the share of transactions that include multiple guns.
Seasonal adjustments are calculated with the U.S. Census Bureau’s X13-ARIMA-SEATS software, and account for factors such as cyclical changes in demand, the number of days in a month, and the timing of holidays. This adjustment allows meaningful comparisons to be made between months.
The NICS data includes four categories of transfer checks: handgun, long gun, multiple, and other. More than one handgun or long gun can be transferred during a single check; “multiple” checks are only conducted when a transaction involves two types of guns. For this reason, we add multiple checks to both the handgun and long gun background check counts. Those totals are then multiplied by 1.1, a “multiple gun sales factor” proposed by Brauer to account for the average number of guns sold per transaction, based on his interviews with firearms retailers.
The transfer data also includes a category for “other” firearms, which includes “frames, receivers, and other firearms that are neither handguns nor long guns.” These checks, which tend to make up 3-4 percent of the annual total, are not included in The Trace’s estimates. Other categories of background checks, such as permit, pawn shop, and administrative, are not used for producing the estimates.
The estimates undercount firearm sales due to state-level permit laws. Some states require prospective handgun buyers to obtain purchase permits, which then exempt them from background checks at the point of sale. Many other states issue some form of permit, such as concealed weapons permits, that exempt bearers from NICS background check requirements. Because some buyers in these states receive permit checks in lieu of transfer checks, our estimates do not account for those sales.
Hawaii, which requires gun buyers to obtain purchase permits, does not report any NICS transfer checks.
Background check data also does not capture private sales, which are estimated to make up 13 percent of all sales, and are not reflected in these estimates.
