Geofence Warrants: An Attack on the Fourth Amendment
A geofence warrant searches a crowd to find a suspect, inverting probable cause and particularity. How Carpenter, the Chatrie circuit split, and the Supreme Court's June 2026 decision settle whether the acquisition is a search — and what the Court left open.
An ordinary warrant names a person or a place and describes what officers expect to find. A geofence warrant does the opposite. It draws a shape on a map, sets a window of time, and asks a provider to return every device that was inside it. The result is a search that begins with a population and narrows toward a suspect, rather than beginning with a suspect. Everyone who happened to be nearby is swept into the first stage of the inquiry.
The Fourth Amendment commands that warrants describe the place to be searched and the things to be seized with particularity, and that they rest on probable cause. A request defined only by a polygon and a clock sits uneasily against both commands. Courts confronting the technique have divided sharply, and in 2026 the question reached the Supreme Court for the first time. What follows traces how the geofence warrant works, why it strains settled doctrine, and where the law now stands.
What a geofence warrant actually asks for
A geofence warrant — sometimes called a reverse-location warrant — does not start with a known suspect whose movements police want to confirm. It starts with a crime scene and a question: which devices were here? To answer it, investigators have historically turned to one repository large enough to be useful, Google’s store of “Location History” data, internally known as Sensorvault. Where conventional location requests target a named account, a geofence request targets coordinates and a clock, and the identities come later.
The investigative appeal is obvious. When there is no eyewitness, no usable surveillance, and no informant, a geofence can surface leads from nothing more than a timestamp and a map. The constitutional problem is equally obvious: the same draw that catches a suspect also catches every commuter, resident, delivery driver, and passerby whose phone was logging location at that moment.
The staged-disclosure process Google used
Google did not simply hand over a list of names. To blunt the privacy intrusion, it adopted a three-step protocol that most geofence litigation now treats as the template. The structure of that protocol is central to the doctrinal fight, because each step exposes a different quantum of data about a different-sized group.
At step one, Google returned anonymized, numbered records for every device whose Location History placed it inside the requested polygon during the requested window. At step two, investigators selected a subset of those numbered devices and asked Google to widen the picture — location points beyond the original geofence and time frame — to distinguish a likely participant from someone merely passing through. At step three, for the few devices that survived that winnowing, police asked Google to deanonymize the records and disclose subscriber identity.
The mechanics are not academic. In the case that reached the Supreme Court, the step-one geofence around a Virginia credit union returned nineteen devices; investigators pulled expanded location data on nine of them at step two; and at step three they sought identifying information for three, one of whom was the defendant, Okello Chatrie.
Defenders of the technique point to step one’s anonymization as proof that no one’s identity is exposed without a narrowing showing. Critics answer that the constitutional event is the seizure and search of the location records themselves — which happens at step one, across the entire population in the box, before any narrowing occurs. Whether the Fourth Amendment attaches at step one or only at step three is the hinge on which much of the litigation turns.
How it inverts particularity and probable cause
A traditional warrant establishes probable cause as to a specific target and then authorizes a search bounded to that target. The geofence runs the logic backward. There is probable cause that someone who was present committed the offense, but no individualized cause as to any of the people the warrant actually reaches at step one. The warrant identifies a place and a time with precision; what it cannot do is particularize the persons or the data to be seized, because identifying them is the entire purpose of the search.
This is the feature that draws comparisons to the general warrants and writs of assistance that the Fourth Amendment was written to forbid. A general warrant licenses a search of everyone in a class in the hope of finding a wrongdoer. A geofence warrant, critics argue, does the same thing with a database instead of a doorway: it authorizes the examination of an undifferentiated group defined only by location, on the theory that the guilty party is somewhere inside it.
Courts that have upheld the technique tend to characterize the intrusion as modest and the staging as a built-in limit. Courts that have condemned it focus on step one, where the records of every innocent bystander are pulled before anyone has shown they are anything but innocent.
The Carpenter backdrop and the third-party doctrine
None of this can be assessed without Carpenter v. United States. There the Supreme Court held that the government conducts a Fourth Amendment search when it acquires historical cell-site location information from a wireless carrier, and that it generally needs a warrant to do so. The decision carved a significant exception into the third-party doctrine — the rule that a person loses Fourth Amendment protection in information voluntarily turned over to a third party such as a bank or phone company.
Carpenter reasoned that comprehensive, automatically generated location data is different in kind from the financial or telephone metadata of earlier cases. Such data offers an “intimate window” into a person’s life, is generated without meaningful choice, and accumulates a record of movement that no traditional surveillance could match. For those reasons the Court declined to treat cell-site records as voluntarily shared in the way the third-party doctrine assumes.
The unresolved question is how far that reasoning travels. Google Location History is, if anything, more precise than the cell-site data in Carpenter — GPS-grade rather than tower-grade. But it is also arguably more “voluntary,” because a user must opt into Location History and can switch it off. That tension drives the entire geofence debate: is location data given to Google more like the bank records the third-party doctrine leaves unprotected, or more like the cell-site records Carpenter shielded? The courts have not agreed.
Carpenter expressly described itself as a “narrow” decision and disclaimed any view of “novel” surveillance techniques. Geofence warrants are precisely such a technique. Lower courts have therefore had to extrapolate from Carpenter‘s logic rather than apply a holding, and they have extrapolated in opposite directions.
United States v. Chatrie: the case that reached the Court
The Virginia prosecution of Okello Chatrie became the leading test. After an armed robbery of a credit union near Richmond in 2019, a detective obtained a geofence warrant for Google Location History within a roughly 150-meter radius of the bank for a one-hour window. The three-step process eventually led to Chatrie, who moved to suppress the evidence.
The district court
The Eastern District of Virginia held that the warrant “plainly” violated the Fourth Amendment: police lacked probable cause as to each of the nineteen device users swept in at step one, and the warrant failed the particularity requirement. The court nonetheless denied suppression, applying the good-faith exception to the exclusionary rule on the ground that the technique was novel and the detective’s reliance on a judicially issued warrant was objectively reasonable.
The Fourth Circuit panel and the en banc reversal
On appeal, a divided three-judge panel of the Fourth Circuit in 2024 reasoned that obtaining two hours of Location History was not a search at all, because the data fell within the third-party doctrine and amounted to a record of a brief trip far less revealing than the long-term tracking in Carpenter. The full court then reheard the case en banc and, on April 30, 2025, affirmed the denial of suppression by a 14–1 vote — but on conspicuously fractured reasoning. The court issued only a brief per curiam judgment; the rationale splintered across multiple separate opinions. On the threshold question of whether a Fourth Amendment search even occurred, the judges divided evenly, seven to seven. The one point a working majority could rest on was the good-faith exception, which made suppression unavailable regardless of the constitutional answer. Judge Gregory dissented, concluding both that a search occurred and that good faith should not rescue it.
Because the en banc court resolved the case on good faith without settling whether geofence warrants are constitutional, it left the central question open — and sharpened a circuit split. The Supreme Court granted certiorari in January 2026, heard argument on April 27, 2026, in Chatrie v. United States, No. 25-112, and decided the case on June 29, 2026. By a five-Justice majority the Court held that police conduct a Fourth Amendment search when they acquire a person’s Location History from Google, because an individual has a reasonable expectation of privacy in his cell-phone location information — an extension of Carpenter to GPS-grade provider data. Justice Kagan wrote for the Court, joined by Chief Justice Roberts and Justices Sotomayor, Kavanaugh, and Jackson; Justice Jackson filed a concurrence joined by Justice Sotomayor; Justice Gorsuch concurred in the judgment; and Justices Alito and Barrett each filed dissents. It was the first time the Court had confronted a geofence warrant directly.
The holding is narrower than it first appears. The Court answered only the threshold question on which it had granted review — whether acquiring the location data is a search — and went no further. Because the Fourth Amendment forbids only unreasonable searches, holding that a search occurred does not by itself decide the case. The Court did not resolve whether the three-step geofence warrant satisfied particularity or established probable cause at each step, and it did not reach the good-faith or suppression questions. Describing itself as “a court of review, not of first view,” it left those issues to the Fourth Circuit on remand. The en banc judgment was vacated and the case remanded, so the suppression outcome for Chatrie himself is not yet final.
A 7–7 split on whether a search occurred is unusual and telling. It means a federal appellate bench, fully briefed, could not agree on the most basic threshold question. The good-faith escape hatch let the court avoid deciding — but good faith is a doctrine about remedies, not rights. It tells officers their evidence is safe; it does not tell anyone whether the search was lawful. The Supreme Court has now begun to close that gap, holding that the geofence acquisition is a search — while leaving whether this particular warrant was reasonable, and whether good faith still shields the evidence, to the lower courts on remand.
The wider docket: Smith, Meza, and the Illinois rejections
Chatrie is not the only voice. While the Fourth Circuit was dodging the merits, the Fifth Circuit confronted them head-on. In United States v. Smith, decided August 9, 2024, the court held that geofence warrants are unconstitutional as “modern-day general warrants” that authorize a search of everyone in a defined area without individualized cause — yet, like the courts before it, declined to suppress, applying the good-faith exception given the novelty of the technique.
State courts have reached comparable conclusions. In People v. Meza, a California Court of Appeal became the first appellate court in the country to rule on a geofence warrant and found it overbroad and insufficiently particular: it authorized identifying anyone within six large Los Angeles search areas, including places and times where the suspects could not have been. The court still affirmed the denial of suppression on good-faith grounds.
At the trial level, magistrate judges in the Northern District of Illinois twice refused to issue geofence warrants in In re Search of Information Stored at Premises Controlled by Google, finding the applications “completely devoid of any meaningful limitation” and unable to establish probable cause to seize the records of users who could not be tied to the offense. Those rulings, rejecting the warrants before any search occurred, were among the earliest signals that the technique would not pass unexamined.
The pattern across these decisions is striking and consistent. Courts increasingly agree that the technique strains or breaks the Fourth Amendment, yet defendants almost never win suppression, because the good-faith exception absorbs the violation. The constitutional doubt is real; the practical remedy has been nearly absent.
The data: how fast geofence requests grew
The stakes are not hypothetical. Google’s own transparency disclosures show how quickly the technique scaled before it drew judicial scrutiny.
Google moves the data on-device
In December 2023 Google announced changes that may render the question partly moot in practice. Going forward, Location History would default to storage on the user’s own device rather than in Google’s cloud; the default retention period would shrink to three months; and cloud backups would be encrypted so that Google itself could not read them. Google indicated it would no longer be able to respond to geofence warrants with the kind of comprehensive Sensorvault sweep that made the technique possible, because it would no longer hold the aggregated data in a form it could query.
That shift matters, but it is not a constitutional answer, and it should not be mistaken for one. It is a corporate policy decision, reversible at will and unaccompanied by any enforceable guarantee. It addresses one provider’s retroactive dragnet, not the underlying legal question, and it leaves untouched the records already collected, the data other providers may hold, and the prospective tracking that a different warrant could still compel. A doctrine that depends on a company’s product roadmap is no doctrine at all. The Supreme Court’s resolution of Chatrie now governs the threshold question wherever the data still exists — the acquisition is a search — even as the reasonableness of any given warrant remains to be worked out case by case.
Reverse keyword warrants: the parallel problem
The geofence warrant has a close cousin. A reverse keyword warrant asks a provider to identify everyone who searched a particular term, phrase, or address within a time window — inverting search the same way a geofence inverts location. The Colorado Supreme Court became the first state high court to address the device in People v. Seymour, where Denver police sought the IP addresses of everyone who had searched a victim’s home address before a fatal arson.
The court declined to issue a broad pronouncement. It recognized a constitutionally protected privacy interest in search history under the state constitution and acknowledged that such warrants implicate expressive freedom, yet it upheld the evidence on good-faith grounds — the now-familiar pattern. A dissent argued the warrant was invalid and that good faith could not save it. The doctrinal parallel to geofence litigation is exact: a recognized intrusion, an unsettled merits question, and a remedy defeated by good faith.
Remedies, suppression, and the good-faith trap
The single most consequential feature of geofence litigation is not any holding on the merits. It is the recurrence of the good-faith exception. From Chatrie to Smith to Meza to Seymour, courts have found constitutional problems and then declined to suppress, reasoning that officers relied reasonably on a novel technique not yet held unlawful. The result is a peculiar equilibrium: the warrants may be unconstitutional, but the evidence comes in anyway.
This creates a structural difficulty for anyone hoping the courts will police the technique. If every geofence warrant is excused as a good-faith reliance on unsettled law, the law never settles through suppression, because no defendant ever obtains the remedy that would force the issue. That is part of why the Supreme Court’s decision in Chatrie matters so much — and part of why it does not end the debate. By holding that the acquisition is a search, the Court removed the threshold novelty that good-faith reasoning had leaned on. But it expressly declined to decide whether the evidence against Chatrie must be suppressed, returning that question to the lower courts; whether good faith continues to absorb the violation now that the search question is settled is the issue the remand will test.
Innocent bystanders and the civil-liberties stakes
Beyond doctrine lies the lived consequence. The defining trait of a geofence warrant is that it reaches people who did nothing. A device logged near a crime scene belongs, far more often than not, to a neighbor, a commuter, a customer, or a passerby. Documented cases include a man identified as a burglary suspect because his fitness app placed him cycling past the scene. The technique converts ordinary presence into investigative suspicion.
The chilling effects compound when location is paired with expression. A geofence drawn around a place of worship, a clinic, a protest, or a political gathering would identify everyone who attended — a capability that touches the First Amendment as much as the Fourth. The same is true of reverse keyword warrants, which can expose who searched a sensitive term. These are not edge cases dreamed up by litigators; they are the direct, foreseeable output of a tool that searches a population to find a person.
Legislative responses
Some lawmakers have not waited for the courts. New York’s proposed Reverse Location Search Prohibition Act would amend the state’s Criminal Procedure Law to bar reverse-location and reverse-keyword searches of people defined only by where they were or what they searched, rather than by individualized suspicion — whether or not a warrant issues. If enacted it would be the first state statute to prohibit the techniques outright, and it has drawn support from a coalition of technology companies.
At the federal level, reform has tended to ride on broader surveillance bills rather than geofence-specific statutes, and none has yet become law. The legislative path offers something the judicial path has struggled to deliver: a categorical rule that does not depend on the good-faith posture of any individual officer. But legislation moves slowly, and for now the governing law remains whatever the courts say it is.
Where this is heading
The Supreme Court has now answered the threshold question. Its decision in Chatrie establishes that acquiring Google Location History is a Fourth Amendment search and that the third-party doctrine does not strip GPS-grade provider data of protection — a reshaping of the technique nationwide, because step one can no longer be treated as no search at all. But the Court stopped there. It did not decide whether the staged-disclosure structure can satisfy particularity, whether the warrant rested on probable cause at each step, or whether the good-faith exception still saves the evidence; those questions return to the Fourth Circuit on remand. The doctrine is thus half-settled: the search question resolved, the reasonableness and remedy questions still open. Either way, the practical landscape has already shifted: Google has moved much of the data beyond easy reach, legislatures are circling, and the reverse-keyword variant is rising in parallel. The constitutional account of searches that begin with a crowd rather than a suspect is no longer hypothetical; its first chapter has now been written, and the rest is being written on remand.
Golden Gate Legal Review offers commentary and analysis on developments like these; it is not legal advice. Continuing coverage appears in the commentary section and the case tracker. Related analysis takes up employee privacy rights while working from home and the strain that the same surveillance pressures place on the third-party doctrine in the cloud era.
Questions readers ask
What is a geofence warrant?
A geofence warrant, also called a reverse-location warrant, asks a technology provider to identify every device recorded within a drawn geographic area during a set time window. Instead of naming a suspect, it names a place and a clock and works backward toward whoever was present.
How is a geofence warrant different from an ordinary warrant?
An ordinary warrant establishes probable cause as to a specific person or place and then authorizes a bounded search. A geofence warrant reverses the order: it identifies an area, sweeps in everyone who was there, and only afterward narrows toward a suspect. The identities are the goal of the search, not its premise.
What is Sensorvault?
Sensorvault is the internal name for Google’s large repository of “Location History” data — the aggregated, time-stamped location records that made geofence warrants possible. When investigators sought devices near a crime scene, this database was queried to produce them.
How does the three-step disclosure process work?
At step one, Google returns anonymized, numbered records for every device inside the geofence during the window. At step two, investigators ask for expanded location data on a selected subset to filter likely participants from passersby. At step three, Google deanonymizes the few remaining devices and discloses subscriber identity.
Why do critics call geofence warrants “general warrants”?
Because they authorize the search of an undifferentiated group defined only by location, without individualized probable cause as to any member. That is the structure the Fourth Amendment was written to prohibit. The Fifth Circuit adopted this characterization in United States v. Smith (2024).
What did Carpenter v. United States decide?
Carpenter, 585 U.S. 296 (2018), held that acquiring historical cell-site location information is a Fourth Amendment search generally requiring a warrant, and that comprehensive location data is not stripped of protection by the third-party doctrine. It described itself as narrow and did not address novel techniques like geofence warrants.
What is the third-party doctrine?
It is the rule that a person generally loses Fourth Amendment protection in information voluntarily disclosed to a third party, such as a bank or telephone company. Carpenter created an exception for comprehensive location data, and whether Google Location History falls within that exception is central to the geofence debate.
What happened in United States v. Chatrie?
A geofence warrant led police to Okello Chatrie after a Virginia robbery. The district court (2022) found the warrant violated the Fourth Amendment but denied suppression on good-faith grounds. The Fourth Circuit ultimately affirmed en banc on April 30, 2025, on fractured reasoning, and the Supreme Court heard argument on April 27, 2026. On June 29, 2026, the Court held that acquiring his Location History was a Fourth Amendment search, vacated the en banc judgment, and remanded; it left the particularity, probable-cause, and suppression questions for the lower courts, so the outcome for Chatrie himself is not yet final.
What did the en banc Fourth Circuit hold in Chatrie?
By a 14–1 vote it affirmed the denial of suppression, but on splintered reasoning. The judges divided seven to seven on whether a Fourth Amendment search even occurred and rested the judgment on the good-faith exception, leaving the constitutional question unresolved until the Supreme Court took up the case and held, in June 2026, that the acquisition was a search.
Did the Supreme Court decide whether geofence warrants are constitutional?
In part. On June 29, 2026, in Chatrie v. United States, No. 25-112, the Court held that acquiring a person’s Google Location History is a Fourth Amendment search, extending Carpenter. But it did not decide whether geofence warrants are constitutional overall: it expressly left open whether the staged warrant satisfied particularity and probable cause, vacated the judgment below, and remanded those questions to the Fourth Circuit.
What did United States v. Smith decide?
The Fifth Circuit held on August 9, 2024, that geofence warrants are unconstitutional as modern-day general warrants, but declined to suppress the evidence under the good-faith exception. The decision conflicts with the Fourth Circuit’s approach and helped create the split the Supreme Court resolved on the search question in June 2026.
What was the holding in People v. Meza?
A California Court of Appeal held in 2023 that a geofence warrant covering six large Los Angeles areas was overbroad and lacked particularity, because it authorized identifying anyone present without individualized cause. It nonetheless upheld the denial of suppression on good-faith grounds.
Why do defendants keep losing even when courts find a violation?
Because of the good-faith exception to the exclusionary rule. Courts repeatedly conclude that officers relied reasonably on a novel technique not yet held unlawful, so evidence is admitted even where the warrant is found unconstitutional. The remedy, not the right, is what fails.
What changed when Google moved Location History on-device in 2023?
In December 2023 Google began storing Location History on users’ devices by default, shortened default retention to three months, and encrypted cloud backups so it cannot read them. Google indicated it could no longer respond to geofence warrants with a Sensorvault-style sweep, because it would no longer hold the aggregated data.
Does Google’s policy change make geofence warrants unconstitutional?
No. It is a corporate decision that limits one provider’s ability to respond, not a constitutional ruling. It is reversible, does not bind other providers, and leaves the underlying legal question for the courts. The constitutional account still depends on cases like Chatrie.
What is a reverse keyword search warrant?
It asks a provider to identify everyone who searched a particular term, phrase, or address within a time window. Like a geofence, it inverts the usual order of a search. The Colorado Supreme Court addressed one in People v. Seymour (2023), recognizing a privacy interest in search history but admitting the evidence on good faith.
Can a geofence warrant identify innocent people?
Yes, and by design it sweeps in everyone present. Devices belonging to neighbors, commuters, and passersby are caught at the first stage. Documented cases include a person treated as a suspect because a fitness app logged him passing the scene, illustrating the technique’s bystander problem.
Are there First Amendment concerns?
Yes. A geofence drawn around a protest, a clinic, or a place of worship would identify everyone who attended, and reverse keyword warrants can reveal who searched sensitive terms. Both implicate expressive and associational freedoms alongside Fourth Amendment privacy.
Are any laws being passed to restrict these warrants?
New York’s proposed Reverse Location Search Prohibition Act would bar reverse-location and reverse-keyword searches of people defined only by location or search terms, and would be the first such state statute if enacted. Federal restrictions have been proposed within broader surveillance bills but have not become law.
Where can someone follow how this area of law develops?
The Supreme Court resolved the threshold search question in Chatrie in June 2026, but the particularity, probable-cause, and suppression questions remain open on remand, so the law is still developing. Ongoing coverage and case updates appear in the commentary section and the case tracker. This material is analysis and commentary, not legal advice.
