DHS Slashes Its Nevada Non-Citizen Voter Estimate from 15,903 to 185 Verified Cases
The Department of Homeland Security has confirmed to Nevada officials that only 185 names on its flagged voter-roll list have been verified as non-citizens — a steep drop from the initial figure of as many as 15,903 that the department presented in July, according to records obtained by the Guardian.
DHS Secretary Markwayne Mullin sent letters to secretaries of state across multiple jurisdictions in mid-July warning about non-citizen registrations. In a letter to Nevada dated 17 July 2026, DHS suggested there may be as many as 15,903 non-citizens registered to vote in the state. The claim drew immediate scrutiny from Nevada election officials, who demanded that DHS explain its methodology and supporting evidence (Democracy Docket). Secretary of State Francisco Aguilar, a Democrat, asked the Trump administration to provide evidence backing the assertion (Nevada Independent).
The 15,903 figure entered the public arena at the highest level. In a primetime address on 16 July 2026, Donald Trump claimed a DHS review had found approximately 278,000 non-citizens registered to vote across four states: Nevada, California, Pennsylvania, and New Jersey. That aggregate figure was subsequently walked back in follow-up letters to all four states, where DHS described the 278,000 as a ceiling of potential non-citizen registrations rather than a confirmed count.
To build its list, DHS compared Nevada's public voter rolls from July 2025 against internal immigration data from US Citizenship and Immigration Services (USCIS). The method works like a cross-reference: it matches names on voter registration lists against entries in a database of people who are not citizens. But a match between those two records does not automatically mean the person was a non-citizen when they registered to vote. Someone may have become a naturalized citizen — meaning they completed the legal process of gaining citizenship — between the time the immigration record was created and the time they registered. Or the match may simply reflect a data discrepancy, such as two people sharing the same name.
In a 14 August 2026 meeting, DHS officials told Nevada election officials that the 15,903 number represented the ceiling of how many non-citizens could be registered, and that the figures given to the public were preliminary. A DHS spokesperson subsequently confirmed that only 185 people on the Nevada list had been verified as non-citizens, describing those 185 as the cases that had gone through manual review (Guardian).
DHS also sent Nevada an additional 6,218 names it had not yet manually reviewed but for which it had made a "higher confidence match" that they could be non-citizens. The status of those names remains unconfirmed pending further review.
Aguilar has stated his office would not risk disenfranchising eligible voters — meaning stripping them of their right to vote — without the highest confidence of their ineligibility. He positioned Nevada's verification process as a guardrail against premature list purges based on unverified database matches.
The broader context here is a familiar pattern in election administration. Automated cross-system matching between voter rolls and immigration databases tends to produce a large pool of potential flags, but manual review typically shrinks that pool dramatically. The gap between preliminary ceilings and confirmed cases reflects the difficulty of reconciling records across systems that track different statuses at different points in time. In Nevada's case, the confirmed count of 185 represents roughly 1.2% of the 15,903 ceiling that DHS initially presented.
The stakes involve both election integrity and voter access. Non-citizen registration is already a federal crime, and election officials across the political spectrum acknowledge the importance of clean rolls. But the distance between DHS's initial public posture and its confirmed findings raises questions about how preliminary data is communicated, particularly when it is amplified at the presidential level before manual verification is complete. The four-state sweep, with its 278,000 aggregate ceiling, now faces the same scrutiny that Nevada's share received: whether the methodology and its limitations were adequately conveyed before the numbers entered public discourse.
