State of Minnesota v. Ivan Contreras-Sanchez

Opinion

STATE OF MINNESOTA

IN SUPREME COURT

A22-1579

Court of Appeals Hennesy, J.

Concurring, Hudson, C.J.

Dissenting, McKeig, Moore, III, JJ.

Took no part, Gaïtas, J.

State of Minnesota, Respondent,

vs. Filed: April 15, 2026 Office of Appellate Courts Ivan Contreras-Sanchez

Appellant.

________________________

Keith Ellison, Attorney General, Saint Paul, Minnesota; and

Mary F. Moriarty, Hennepin County Attorney, Adam E. Petras, Senior Assistant County Attorney, Minneapolis, Minnesota, for respondent.

Cathryn Middlebrook, Chief Appellate Public Defender, Jennifer Workman Jesness, Assistant State Public Defender, Saint Paul, Minnesota, for appellant.

Teresa Nelson, Alicia Granse, American Civil Liberties Union of Minnesota, Minneapolis, Minnesota;

Jennifer Stisa Granick, American Civil Liberties Union Foundation, San Francisco, California; and

Brett Max Kaufman, American Civil Liberties Union Foundation, New York, New York, for amici curiae American Civil Liberties Union and American Civil Liberties Union of Minnesota.

2

Leita Walker, Ballard Spahr, LLP, Minneapolis, Minnesota, for amicus curiae Electronic Frontier Foundation.

Shauna Faye Kieffer, Minnesota Association of Criminal Defense Lawyers, Minneapolis, Minnesota, for amici curiae National Association of Criminal Defense Lawyers and Minnesota Association of Criminal Defense Lawyers.

________________________

SYLLABUS

1. Because cell phone users have a reasonable expectation of privacy in their location data stored by Google, the government conducted a search under Article I, Section 10, of the Minnesota Constitution when it accessed a cell phone user’s location data stored by Google.

2. Geofence warrants are not categorically prohibited general warrants under the Minnesota Constitution.

3. Because the warrant application in this case established a fair probability that Google’s servers would contain evidence of a crime, there was probable cause to issue the geofence warrant; this geofence warrant did not require a probable cause nexus for every person within the geofence.

4. A geofence warrant is insufficiently particular under the Minnesota Constitution when it gives law enforcement unchecked discretion to determine which device identification numbers inside the geofence will be subject to an additional search for more location information.

Reversed and remanded.

3

OPINION

HENNESY, Justice.

In this case we are asked for the first time to consider the constitutionality of a geofence warrant, a type of warrant that allows law enforcement to obtain information from a tech company regarding the presence of customer devices, including cell phones, located within selected geographic coordinates. Appellant Ivan Contreras-Sanchez was convicted of second-degree intentional murder for killing Manuel Mandujano. Law enforcement connected Contreras-Sanchez to the crime using data gathered under a geofence warrant. Contreras-Sanchez appealed his conviction, arguing that the geofence warrant violated the United States and Minnesota Constitutions. The court of appeals affirmed his conviction.

We conclude that the Minnesota Constitution does not provide the government unrestricted access to a person’s location data stored by Google and that the warrant here failed to satisfy the particularity requirement. Because we conclude that Contreras- Sanchez’s rights under the Minnesota Constitution were violated, we have no need to address whether this warrant satisfied the requirements under the Fourth Amendment to the United States Constitution. Specifically, we conclude that the government conducted a search under the Minnesota Constitution when it accessed Contreras-Sanchez’s location data stored by Google, which holds information about intimate and deeply private aspects of a person’s life: their familial, religious, political, health care, and other highly sensitive activities and relationships. We acknowledge, however, that geofence warrants are a useful tool for law enforcement and reject the notion that geofence warrants are per se unconstitutional general warrants. We also conclude that the geofence warrant here was

4

supported by probable cause. Finally, we hold that the geofence warrant in this case was not constitutionally particular, because it did not require a judge to assess law enforcement’s decision to use data from a narrowly defined (both in time and geography) geofence warrant to significantly expand the warrant’s geographic and temporal scope. Had there been judicial review of the reasonableness of that expanded search, the result here may have been different.

Because the court of appeals held the geofence warrant was sufficiently particular, it did not decide whether the good-faith exception to the exclusionary rule applies in this context or whether the error was harmless. The parties did not ask for review of these issues in their petition to our court. Consequently, we reverse the court of appeals and remand to that court to consider the merits of issues it did not reach. 1 FACTS

On April 26, 2021, police responded to a report of a body found in a drainage culvert in Castle Rock Township. A forensic examiner identified the body as Manuel Mandujano, who had last had contact with family on March 25, 2021. Officers applied for a geofence warrant, requesting location history data generated by devices registering within a specified geographic location near the body and stored by Google. The officers asked for this data from the date Mandujano disappeared until the date his body was found. The geofence

1 As we note below, because the issue of whether the execution of a geofence warrant violates the Fourth Amendment is currently pending before the United States Supreme Court in Chatrie v. United States, 136 F.4th 100 (4th Cir. 2025), cert. granted, No. 25-112, 2026 WL 120676 (U.S. Jan. 16, 2026), there may be circumstances in which it is appropriate for the court of appeals to consider the Supreme Court’s decision in that case on remand.

5

specified in the warrant request had a rectangular perimeter of 65 feet by 290 feet and encompassed the culvert, the public road nearby, and a portion of the right-of-way ditch. According to the warrant application, only “a small number” of vehicles use the road and there are only three residences on that section of road.

The warrant application stated that a medical examiner had determined that Mandujano’s “manner of death was [h]omicide, however a cause of death was still unknown.” According to the application, an informant told police they knew someone who had been involved in Mandujano’s murder and that others had also been involved. The informant said that the people involved in the murder had cell phones, though the informant was unsure which brand and which providers the participants used. The warrant application identified the place to be searched as “Google, LLC.” Regarding Google’s data collection processes, the application stated: “Google, LLC retains and uses location information for individuals who use a wide range of Google productions [sic]. This information kept by Google also contains location information, which can be very accurate.”

The warrant application outlined a three-step process law enforcement would use to obtain data from Google. In step one, the application requested that Google be ordered to provide anonymous device identification numbers (device IDs) for every data point within the geofence during the one-month period from March 25, 2021, when Mandujano went missing, to April 26, 2021, when his body was found. An officer attested that there “is no way . . . to identify the related user of the device with this information alone.”

6

In step two, the warrant application provided:

• Law enforcement will analyze this location data to identify users who may have witnessed or participated in the [s]ubject [o]ffenses . . . .

• For those accounts identified as relevant to the ongoing investigation through an analysis of provided records, and upon demand, [Google]

shall provide additional location history outside of the predefined area for those relevant accounts to determine path of travel. This additional location history shall not exceed 60 minutes plus or minus the first and last timestamp associated with the account in the initial dataset. (The purpose of path of travel/contextual location points is to eliminate outlier points where, from the surrounding data, it becomes clear the reported point(s) are not indicative of the device actually being within the scope of the warrant.)

(Emphasis added.) In other words, step two had two phases: first, it authorized officers to determine which device IDs merited further search for additional location history outside the targeted geofence boundary for 60 minutes before and after those device IDs were within the geofence boundary, to show their paths of travel; and second, upon officers’ “demand,” it compelled Google to release the additional location history as to the selected device IDs. At step three, the warrant application authorized officers to analyze the step- two data and “demand” that Google identify the subscribers connected to relevant accounts. 2

2 Notably, the warrant application and the proposed warrant used different language to describe step three. The warrant application stated that police would apply for a second warrant using the data collected from steps one and two and laying out “the probable cause related to the specific ‘Device ID(s)’ that were discovered as related to the crime” to obtain the subscriber information associated with relevant device IDs. The language in the proposed warrant itself simply required Google to provide law enforcement with subscriber information for “accounts identified as relevant . . . upon demand.” The proposed warrant, whose language the district court adopted in the signed warrant, did not require law enforcement to obtain a second warrant before demanding Google provide subscriber information.

7

A district court judge signed a warrant with language identical to the proposed warrant. Shortly after law enforcement served the warrant, Google informed officers that processing the full month of data requested, from March 25 to April 26, 2021, would “overload [Google’s] systems.” Google recommended breaking this timeframe into shorter portions to process the data faster. Police narrowed their search timeframe and requested data from two seven-day periods: March 25–31 and April 1–7.

Google provided, and police analyzed, this data. Between March 25 and March 31, the data showed that 12 devices entered the geofence. Eleven devices registered locations only once. The data showed that one device was in the area for 10 minutes, registering locations within the geofence 46 times. Plotting these points on a map, officers saw that device had been located directly over the culvert. Data from the second seven-day timeframe revealed an additional 19 devices that registered locations within the geofence, none staying for longer than a few minutes.

Police next asked Google for step-two location data for the device ID that logged 46 data points in the culvert. They did not seek an additional warrant prior to making this request. Step-two data revealed that the device ID registered a location at a gas station during the 60 minutes prior to entering the geofence. Law enforcement reviewed surveillance video from the relevant time at that gas station and saw “four persons of interest” enter the station. An officer recognized one individual and a vehicle from the ongoing investigation.

Police then executed step three in a manner consistent with the warrant application.

Officers applied for a second warrant based on the data they had received and analyzed in

8

the previous two steps and the surveillance video. They narrowed their search to the single device ID that generated 46 data points in the culvert and requested that Google identify all “subscriber and/or user information, including the name, addresses, account status, all email addresses, alternative email addresses, cellular, home and business addresses, and phone numbers” associated with that device ID. A district court judge approved this warrant. Google identified the device ID as registered to Contreras-Sanchez.

In an interview with police, Contreras-Sanchez admitted to participating in Mandujano’s murder. On November 5, 2021, the State charged Contreras-Sanchez with second-degree intentional murder and second-degree unintentional felony murder in connection with Mandujano’s death.

Prior to trial, Contreras-Sanchez moved to suppress all evidence obtained through the geofence warrant. He argued that all geofence warrants are per se unconstitutional general warrants and, in the alternative, that this geofence warrant lacked probable cause, was not sufficiently particular, and was overbroad. The State argued that there is no reasonable expectation of privacy in location data stored at Google. In the alternative, the State argued that this geofence warrant was supported by probable cause and was sufficiently particular. The district court denied the motion to suppress. A jury found Contreras-Sanchez guilty of both counts of second-degree murder. 3 Contreras-Sanchez appealed, arguing to the court of appeals that all geofence warrants are unconstitutional general warrants and the warrant in this case violated the

3 The district court judge only entered conviction on the second-degree intentional murder verdict.

9

Fourth Amendment to the United States Constitution and Article I, Section 10, of the Minnesota Constitution. Specifically, he argued that there was no probable cause to issue this warrant, and that it was not sufficiently particular and was overbroad. The State argued that the geofence warrant was constitutional, but, if it was not, the good-faith exception to the exclusionary rule should apply and the alleged error was harmless. The court of appeals affirmed the district court’s denial of Contreras-Sanchez’s motion to suppress, concluding that the geofence warrant did not violate the United States or Minnesota Constitutions because it was not an unconstitutional general warrant, was supported by probable cause, and was sufficiently particular and not overbroad. State v. Contreras-Sanchez, 5 N.W.3d 151, 171 (Minn. App. 2024). Consequently, the court of appeals did not reach the State’s arguments regarding the good-faith exception and harmless error.

We granted review to address the constitutionality of geofence warrants.

ANALYSIS

We begin our analysis by determining whether law enforcement may access location data stored by Google without a warrant and, if not, determining whether all geofence warrants are categorically prohibited as general warrants. We then consider the characteristics of the warrant application in this case to determine whether there was probable cause to issue the warrant and, finally, whether the warrant was sufficiently particular.

In conducting this analysis, we recognize that neither we nor the United States Supreme Court has previously addressed whether a geofence warrant violates the constitution. We further acknowledge that the United States Supreme Court has recently

10

granted review on related issues. See Chatrie v. United States, 136 F.4th 100 (4th Cir. 2025), cert. granted, No. 25-112, 2026 WL 120676 (U.S. Jan. 16, 2026). We issue our opinion without waiting for the Supreme Court’s disposition so as to avoid further delaying a decision in light of our responsibility to decide the issues presented not only for the parties but for the people of Minnesota. Nor is there any need to stay this case when we have repeatedly held, “without a doubt[,] that Article I, Section 10 of the Minnesota Constitution provides greater protection against suspicionless law enforcement conduct than the Fourth Amendment to the United States Constitution.” State v. Leonard, 943 N.W.2d 149, 156 (Minn. 2020) (citing State v. Carter, 697 N.W.2d 199 (Minn. 2005), and Ascher v. Comm’r of Pub. Safety, 519 N.W.2d 183 (Minn. 1994)). Thus, irrespective of what the Supreme Court decides the United States Constitution permits in Chatrie, we decide this case under the Minnesota Constitution, “[e]xercising our independent authority to interpret our own state constitution to protect the rights of the citizens of Minnesota.” Ascher, 519 N.W.2d at 187 (citing Michigan v. Long, 463 U.S. 1032, 1041 (1983)); see also Long, 463 U.S. at 1040 (“Respect for the independence of state courts, as well as avoidance of rendering advisory opinions, have been the cornerstones of [the United States Supreme Court’s] refusal to decide cases where there is an adequate and independent state ground.”).

We hold, first, that under the Minnesota Constitution there is a reasonable expectation of privacy in a cell phone user’s location data stored by Google. Consequently, law enforcement conducts a search when accessing this data and must obtain a valid warrant to do so. Second, we hold that geofence warrants are not categorically prohibited as general warrants. Third, we conclude that the warrant application in this case established

11

a fair probability that Google’s servers would contain evidence of a crime such that there was probable cause to issue the geofence warrant, and this warrant did not require a probable cause nexus for every person within the geofence. Finally, we conclude the warrant here was insufficiently particular because it allowed police unchecked discretion to determine which device IDs inside the geofence they would subject to an expanded search for additional location information. Because we hold that Contreras-Sanchez’s rights were violated under the Minnesota Constitution, we do not need to reach his Fourth Amendment challenge. See Leonard, 943 N.W.2d at 152 n.1; State v. Flowers, 734 N.W.2d 239, 258 (Minn. 2007). Based on our holdings, we remand to the court of appeals to determine whether the good-faith exception to the exclusionary rule applies or whether the error was harmless. 4 I.

In response to Contreras-Sanchez’s claim that all geofence warrants are general warrants, the State argues that no warrant is necessary for the government to obtain Google user location data. The State asserts that obtaining such user data is not a search because there is no reasonable expectation of privacy in the location data geofence warrants target. We address, first, the threshold question of whether law enforcement’s request for location data within a geofence constitutes a search. For the reasons below, we conclude that when

4 Although we do not need to address Contreras-Sanchez’s Fourth Amendment challenge because we hold that his rights were violated under the Minnesota Constitution, if the Supreme Court concludes in Chatrie that the execution of the geofence warrant violated the Fourth Amendment, we permit the court of appeals on remand to address Chatrie’s applicability to Contreras-Sanchez’s Fourth Amendment arguments.

12

law enforcement accesses such data, it constitutes a search under Article I, Section 10, of the Minnesota Constitution.

A.

Contreras-Sanchez claims that law enforcement conducted an unconstitutional search when it executed the geofence warrant for his location data stored by Google. We conduct a de novo review of “the district court’s legal conclusions on the constitutionality of searches and seizures.” State v. Westrom, 6 N.W.3d 145, 153 (Minn. 2024).

Article I, Section 10, of the Minnesota Constitution protects the “right of the people to be secure in their persons, houses, papers, and effects against unreasonable searches and seizures.” Minn. Const. art. I, § 10. “A search occurs when the government intrudes upon a ‘reasonable expectation of privacy’ to gain information.” Westrom, 6 N.W.3d at 153.

Although we have not addressed whether there is a reasonable expectation of privacy in location data held by Google, we have recognized that “general searches of electronic data threaten individual privacy.” State v. Sardina-Padilla, 7 N.W.3d 585, 599 (Minn. 2024). And we have noted our concern that cell phones can track a person’s location history with alarming accuracy, revealing not only the details of a person’s movements, “but through them his familial, political, professional, religious, and sexual associations.” In re B.H., 946 N.W.2d 860, 869 (Minn. 2020) (citation omitted). In State v. Leonard, we recognized that the Minnesota Constitution affords Minnesotans a reasonable expectation of privacy in the “highly sensitive location information found in a [hotel] guest registry.” 943 N.W.2d at 156. In that case, law enforcement inspected—without a warrant—a registry that state law required hotels to keep recording guests’ names, addresses, and

13

vehicle information, and the names and addresses of any traveling companions. Id. at 153. In holding that this inspection constituted a search, we observed that “most Minnesotans would be surprised and alarmed if th[is] sensitive location information” was available to the government without any particularized suspicion of criminal activity. Id. at 158. We focused on the nature of the private information the search revealed; namely, “highly sensitive location information” that might reveal “deeply private[] activities,” such as whether a guest had “stayed overnight at the hotel to attend a political or religious conference in the hotel ballroom” or had “stayed overnight before a medical appointment in hopes of keeping a diagnosis private.” Id. at 157–59.

In our view, the location data that Google stores presents even greater privacy concerns than those we considered in Leonard. Similar to a hotel registry, the location information Google gathers and stores through cell phones 5 contains highly sensitive information about deeply personal activities. But while a hotel registry records a snapshot in time, the location data stored by Google is retrospective; its servers hold a comprehensive chronology of a person’s movements, tracked by their cell phone. And though geofence warrants may authorize law enforcement to retrieve data from only a brief time span, much like a hotel registry, this short time period may still reveal the most private

5 While the geofence warrant in this case requested data not only for cell phones but all Google-connected devices, we focus here on the location data collected by cell phones for two reasons: first, the relevant device in this case is Contreras-Sanchez’s cell phone, and second, because “cell phones and the services they provide are such a pervasive and insistent part of daily life that carrying one is indispensable to participation in modern society.” Carpenter v. United States, 585 U.S. 296, 315 (2018) (citation omitted) (internal quotation marks omitted).

14

moments of a person’s life, including their “familial, political, professional, religious, and sexual associations.” B.H., 946 N.W.2d at 869 (citation omitted). In the present case, step- two data revealed that the device ID of interest stopped at a gas station prior to entering the geofence. Under other circumstances, this data could easily have revealed a visit to a health care provider, lawyer, gun store, civic organization, political group, charity, or place of worship.

The government’s capacity to obtain, through Google, a comprehensive catalog of a person’s physical movements recorded by the phone they carry is a privacy concern far beyond that posed by a name and address listed on a hotel registry—information in which we have already recognized an expectation of privacy under the state constitution. See Leonard, 943 N.W.2d at 158. We conclude that under Article I, Section 10, of the Minnesota Constitution, that same expectation of privacy applies to the location data Google collects through tracking its users’ cell phones. 6 Importantly, this does not mean

6 The dissent suggests that Contreras-Sanchez did not meet his burden of establishing that he has a subjective expectation of privacy in his own location data stored by Google. The State did not make this argument. Nonetheless, we observe that Leonard does not support the dissent’s implication that Contreras-Sanchez cannot have an expectation of privacy in his sensitive location data without first establishing that he personally took steps to protect his data from the third party—Google. We did not require that Leonard show any subjective expectation of privacy in the information he provided in his particular hotel registry when we concluded that a guest has a reasonable expectation of privacy, under the Minnesota Constitution, in the location information recorded in a hotel registry. 943 N.W.2d at 158–59. In reaching this conclusion, we observed that “most Minnesotans would be surprised and alarmed if the sensitive location information found in the guest registries at hotels, motels, or RV campsites was readily available” to the government “without any particularized suspicion of criminal activity.” Id. at 158. Here, Contreras- Sanchez is the Google user whose records were searched, and he has shown the sensitive location information at issue was his and that his constitutional rights are at stake.

15

the government cannot access users’ data—it only means that the government must obtain a warrant to do so. See Westrom, 6 N.W.3d at 153 (“A search occurs when the government intrudes upon a reasonable expectation of privacy to gain information.”) (citation omitted) (internal quotation marks omitted); State v. Johnson, 813 N.W.2d 1, 5 (Minn. 2012) (observing that whether a search is constitutionally reasonable generally depends on whether the government complies with the warrant clause “by obtaining a warrant from a neutral magistrate based upon probable cause”).

B.

We reach this conclusion by interpreting Article I, Section 10, of the Minnesota Constitution independently of the United States Supreme Court’s articulation of Fourth Amendment protections. See State v. Fuller, 374 N.W.2d 722, 726 (Minn. 1985) (observing that “as the highest court of this state, we are independently responsible for safeguarding the rights of [our] citizens” (alternation in original) (citation omitted) (internal quotation marks omitted)). We observe, however, that when we interpret our constitution independently of the Supreme Court’s Fourth Amendment jurisprudence, that court’s decisions “interpreting a comparable provision of the federal constitution that, as here, is textually identical to a provision of our constitution, [are] of inherently persuasive, although not necessarily compelling, force.” Id. at 727. Here, the Supreme Court’s analysis of expectations of privacy under the Fourth Amendment in the context of similar technology, to date, supports—although it does not compel—our analysis.

Similar to Article I, Section 10, the Fourth Amendment guarantees people the right “to be secure in their persons, houses, papers, and effects, against unreasonable searches

16

and seizures.” U.S. Const. amend. IV. “[T]he Fourth Amendment protects people, not places.” Katz v. United States, 389 U.S. 347, 351 (1967). And it protects not only “property rights,” but “certain expectations of privacy as well.” Carpenter v. United States, 585 U.S. 296, 304 (2018) (citation omitted). A search occurs “[w]hen an individual seeks to preserve something as private, and his expectation of privacy is one that society is prepared to recognize as reasonable.” Id. (citation omitted) (internal quotation marks omitted).

The Fourth Amendment aims to secure “the privacies of life” against “arbitrary power.” Boyd v. United States, 116 U.S. 616, 630 (1886). “A central aim of the Framers was ‘to place obstacles in the way of a too permeating police surveillance.’ ” Carpenter, 585 U.S. at 305 (quoting United States v. Di Re, 332 U.S. 581, 595 (1948)). The United States Supreme Court has applied these principles to protect individuals from being “at the mercy of advancing technology.” Id. (citation omitted).

In Carpenter v. United States, the Supreme Court held that a person has a reasonable expectation of privacy in cell-site location information (CSLI), data similar to the location data at issue in this case. Id. at 313. Cell phones automatically create CSLI, which third parties—specifically, wireless carriers—store. 7 Id. at 300–01. In analyzing whether cell phone users have a reasonable expectation of privacy in CSLI, the Supreme Court noted that “a majority of this Court has already recognized that individuals have a reasonable expectation of privacy in the whole of their physical movements” in United States v. Jones,

7 Communication between cell phones and cell towers creates CSLI. “Cell phones continuously scan their environment looking for the best signal, which generally comes from the closest cell site. . . . Each time the phone connects to a cell site, it generates a time-stamped record known as [CSLI].” Carpenter, 585 U.S. at 300–01.

17

in which the Supreme Court held that attaching a Global-Positioning-System (GPS) tracking device to a car and monitoring the car’s movement was a search. 8 Id. at 310 (citing Jones, 565 U.S. 400, 415, 430 (2012)). According to the Court, CSLI raises similar concerns with respect to a person’s expectation of privacy in their physical movements, because the data “provides an intimate window into a person’s life, revealing not only his particular movements, but through them his ‘familial, political, professional, religious, and sexual associations.’ ” Id. at 311 (quoting Jones, 565 U.S. at 415 (Sotomayor, J., concurring)). The Court recognized that allowing law enforcement unfettered access to this data contravenes “society’s expectation . . . that law enforcement agents and others would not—and indeed, in the main, simply could not—secretly monitor and catalogue every single movement of an individual’s car for a very long period.” Id. at 310 (quoting Jones, 565 U.S. at 429 (Alito, J., concurring)).

The Court concluded in Carpenter that the “privacy concerns” CSLI presents are more significant than those GPS raises in two respects. Id. at 311. First, CSLI tracks people through the phones they carry, and people carry their phones with them every moment of every day. Id. (observing that “[w]hile individuals regularly leave their vehicles, they compulsively carry cell phones with them” everywhere and always). CSLI thus “follows its owner beyond public thoroughfares and into private residences, doctor’s offices, political headquarters, and other potentially revealing locales.” Id. Second, the

8 In Jones, the Supreme Court held, under a property-rights framework, that a search occurred when the government installed a GPS device on a car and used it to monitor the car’s movement on public streets. 565 U.S. 400, 404–06 (2012).

18

data is “retrospective,” allowing officers to “travel back in time to retrace a person’s whereabouts, subject only to the retention polices of the wireless carriers.” Id. at 312 (observing that GPS data, in contrast, requires police to "know in advance whether they want to follow a particular individual, or when”). This enables law enforcement to use CSLI to track anyone at any time through their cell phone, “not just those belonging to persons who might happen to come under investigation.” Id.

Carpenter and Jones provide additional support for our conclusion that the people of Minnesota have a reasonable expectation of privacy in their cell phone location data stored by Google. Like CSLI, this data tracks a person’s movements through the phone (or other Google-connected device) they carry, providing a near-perfect record of the person’s location. And this location data is more precise than the location data either CSLI or GPS provides. See United States v. Smith, 110 F.4th 817, 833 (5th Cir. 2024) (“Geofences also exemplify the [Supreme] Court’s concern with pinpoint location data—this technology provides more precise location data than either CSLI or GPS.” (citing Geofence Warrants and the Fourth Amendment, 134 Harv. L. Rev. 2508, 2510 (2021)), cert. denied sub nom., Smith v. United States, 146 S. Ct. 356 (mem.) (2025). The location data stored by Google does not rely on a single input from a GPS signal or cell tower, but instead “applies a highly sophisticated synthesis of multiple inputs to pinpoint a mobile device’s exact location.” Barbara Bathke, Google and the Role of Surveillance Intermediaries in Geofence Warrants, 26 Tul. J. Tech. & Intell. Prop. 111, 116 (2024).

While we conclude there is persuasive support in Carpenter and Jones for our decision that Minnesotans have a reasonable expectation of privacy in the sensitive,

19

personal, and comprehensive location data Google collects, in the end, we decide this issue under the Minnesota Constitution alone.

C.

The State advances two arguments in support of its position that cell phone users do not have a reasonable expectation of privacy in location data stored by Google. First, the State contends that users lose any reasonable expectation of privacy when they knowingly expose their location data to a third party. Second, the State argues that there could be no reasonable expectation of privacy in this case because in steps one and two of law enforcement’s warrant application the data requested was anonymous. We address each argument in turn.

1.

Under the third-party doctrine, a defendant loses a reasonable expectation of privacy in information they disclose to a third party. Leonard, 943 N.W.2d at 158. In Leonard, we expressly declined to broadly apply the third-party doctrine to Article I, Section 10, of the Minnesota Constitution. See id. at 158 n.14 (noting that “[w]e have not broadly applied the third-party doctrine to Article I, Section 10 challenges” and that “we do not decide whether [the third-party doctrine] should apply broadly under Minnesota law more generally”). But, as in Leonard, even if we were to generally recognize the third-party doctrine under the Minnesota Constitution, “the principles underlying the third-party doctrine are not implicated in this context”—where the information disclosed to the third party is “sensitive location information.” Id. at 159.

20

In Leonard, we explained that “[a]lthough a guest registry includes seemingly public information—a name and address—the act of recording this information in the guest registry creates sensitive location information because a hotel is a place where people engage in a variety of . . . legal, but often deeply private, activities.” Id. In declining to apply the third-party doctrine, we noted that Leonard’s voluntary disclosure of this information to the hotel “transformed his otherwise public data into sensitive location information that society is prepared to recognize as a reasonable privacy interest.” Id. at 158–59.

As discussed above, the location information Google maintains is far more pervasive, personal, and revealing than the location information at issue in Leonard. For these reasons, we conclude that Minnesotans reasonably expect that the government does not have unrestricted access to users’ location data that Google stores and which holds information about intimate and deeply private aspects of their lives. We also conclude, under Article I, Section 10, of the Minnesota Constitution, that the third-party doctrine does not apply to the location data stored by Google that geofence warrants target. Law enforcement’s execution of the geofence warrant here was a search.

Our approach to the third-party doctrine under the Minnesota Constitution here and in Leonard is similar to, and finds additional persuasive support in, the United States Supreme Court’s approach in Carpenter. In Carpenter, the Court declined to extend the third-party doctrine to CSLI even though cell phone users share their locations with their wireless carriers. 585 U.S. at 308–10. The Carpenter Court noted that the government’s request to exclude “the exhaustive chronicle of location information casually collected by

21

wireless carriers” from Fourth Amendment protection represented “a significant extension of [the third-party doctrine] to a distinct category of information.” Id. at 314. And while the Court acknowledged that a person may have a reduced expectation of privacy in information knowingly shared with a third party, it confirmed that a legitimate expectation of privacy may yet remain, particularly with respect to location information—information for which the Court has shown “special solicitude” in the third-party context. Id. at 314–15.

In deciding whether users maintain any expectation of privacy in CSLI data despite having shared this information with wireless carriers, the Carpenter Court analyzed the “nature of the particular documents sought” and whether users voluntarily provided the information to the third party. Id. at 314, 315. The Court pointed out that in the seminal third-party doctrine cases Smith v. Maryland, 442 U.S. 735 (1979), and United States v. Miller, 425 U.S. 435 (1976), the reason there was no legitimate expectation of privacy in the checks and pen registers 9 at issue was not solely because the defendants had shared them with third parties, but also because the data revealed little in the way of identifying information. Carpenter, 585 U.S. at 314. CSLI, the Court reasoned, is in a different category of information entirely, as it provides a “detailed chronicle of a person’s physical presence compiled every day, every moment, over several years” and so “implicates privacy concerns far beyond those considered in Smith and Miller.” Id. at 315.

9 The Supreme Court described a pen register as “a mechanical device that records the numbers dialed on a telephone by monitoring the electrical impulses caused when the dial on the telephone is released.” Smith, 442 U.S. at 736 n.1 (citation omitted).

22

The Court also rejected the government’s argument that the third-party doctrine must apply to CSLI because cell phone users voluntarily share their location information with third parties. Id. The Court expressed skepticism that a cell phone user truly voluntarily shares this information simply by carrying a phone. Id. Because a cell phone is “indispensable” to modern life and creates CSLI without any affirmative act on the user’s part, the Court concluded that users do not forfeit their reasonable expectation of privacy in that location data by allowing their wireless carriers to generate and store it. Id. at 315–16.

Here, as in Leonard and Carpenter, the State’s request to extend the third-party doctrine to location data stored at Google is not a request “for a straightforward application of the third-party doctrine, but instead a significant extension of it to a distinct category of information.” Id. at 314; see Leonard, 943 N.W.2d at 158–59 (concluding the court of appeals erred in applying the third-party doctrine because the doctrine does not apply to “seemingly public information” that, when shared, becomes sensitive location information). The “nature of the particular documents sought” here is far more like CSLI than a pen register or check. See Carpenter, 585 U.S. at 314. Like a search for CSLI— and unlike a request for the limited information provided by checks or pen registers—a search for location data within a geofence involves rummaging through the archives of a third party with the power to protect or reveal highly personal information based on the history of a cell phone user’s precise physical movements.

Nor does it appear that a cell phone user is knowingly revealing location data to third parties. See id. (“The third-party doctrine partly stems from the notion that an

23

individual has a reduced expectation of privacy in information knowingly shared with another.”). The warrant application in this case noted that “Google, LLC retains and uses location information for individuals who use a wide range of Google productions [sic].” The Carpenter Court concluded that cell phone users do not truly share CSLI because “cell phones and the services they provide are such a pervasive and insistent part of daily life that carrying one is indispensable to participation in modern society.” Id. at 315 (citation omitted) (internal quotation marks omitted). Google products fall within the category of “services [cell phones] provide,” id., and are similarly pervasive and fundamental to participating in our contemporary world. And although Google users may technically “opt in” to location information storage, their understanding of Google’s process of gathering location data is imperfect and users’ choices are hardly informed; any opt in by a user appears, ultimately, illusory. Other courts have noted that Google’s requests for users to share location data are framed innocuously as features that optimize and personalize Google’s service; Google does not explain to users that by opting into sharing location data, Google will continuously track and store users’ physical locations every moment for the foreseeable future and may share them with the government. See Smith, 110 F.4th at 835; United States v. Chatrie, 590 F. Supp. 3d 901, 911–12 (E.D. Va. 2022) (referencing an opt- in prompt that the defendant might have seen stating, in part, “Google needs to periodically store [your] location to improve route recommendations, search suggestions, and more”), aff’d on other grounds, 107 F.4th 319 (4th Cir. 2024), aff’d en banc, 136 F.4th 100 (4th Cir. 2025), cert. granted, No. 25-112, 2026 WL 120676 (U.S. Jan. 16, 2026). Google can, in fact, track and store users’ device location data even when users opt out of location history

24

and other location tracking. In re Search of Info. that is Stored at Premises Controlled by Google LLC (Info. Stored at Google), 579 F. Supp. 3d 62, 70 (D.D.C. 2021). And the process of opting out of location tracking can be so confusing that some of Google’s own employees have indicated in internal emails that they are unsure how to do it and find the process “enough to confuse a privacy focused [software engineer].” Haley Amster & Brett Diehl, Against Geofences, 74 Stan. L. Rev. 385, 396–97 (2022) (alteration in original). For users, “manually deactivating all [location history] sharing remains difficult and discouraged” by Google). Id. It is apparent that users do not assume the full risk of permanently disclosing the entirety of their location history when they use Google’s products or even opt in to Google’s location requests. Instead, the decision of when and to whom to disclose the comprehensive and intimate details of users’ location history that a corporation maintains is left in corporate hands.

We conclude that the third-party doctrine does not apply to the location data Google stores. Although we find additional support for this conclusion in Fourth Amendment jurisprudence, we decide independently under Article I, Section 10, of the Minnesota Constitution that, because users do not knowingly expose the comprehensive chronology of their physical movements by carrying cell phones with Google products that track their location, they do not lose their reasonable expectation of privacy in this data by doing so. 10

10 There is no reason for us to wait for the United States Supreme Court to decide Chatrie under the Fourth Amendment when we have the independent power—and the obligation—to decide this case today under our own constitution. And although the dissent claims we do not decide this case on “adequate and independent state grounds” because we discuss Carpenter and other federal cases, we need not ignore all federal caselaw in order

25

2.

The State’s second argument is that there is no reasonable expectation of privacy in location data obtained in steps one and two of the geofence warrant at issue here because that data is anonymous. But we conclude, consistent with our prior decisions under Article I, Section 10, of the Minnesota Constitution, that the constitutional balance must be struck in favor of protecting the public from even such a minimally intrusive anonymous search. Cf. Ascher, 519 N.W.2d at 187 (“[W]e conclude that the constitutional balance must be struck in favor of protecting the traveling public from even the ‘minimally intrusive’ seizures which occur at a sobriety checkpoint.”).

We do so, in part, based on our skepticism that this location data is truly anonymous, driven by the concern that it is possible to identify users from this location data via cross- reference. 11 One need change the facts of the present case only slightly to illustrate this

to decide an issue on adequate and independent state grounds. See Long, 463 U.S. at 1040– 41 (“If a state court chooses merely to rely on federal precedents as it would on the precedents of all other jurisdictions, then it need only make clear by a plain statement in its judgment or opinion that the federal cases are being used only for the purpose of guidance, and do not themselves compel the result that the court has reached.”). We decide today that there is a reasonable expectation of privacy in this location data not because federal law (or Carpenter) requires us to do so but because our state constitution and jurisprudence require it. Our discussion of Carpenter is appropriate because the parties have briefed and argued it, and it informs our discussion of the third-party doctrine. Carpenter does not, however, compel the result we reach here. 11 We are not alone in expressing this skepticism. The Fifth Circuit, in Smith, suggested that data’s anonymity at the start of the geofence warrant process does not ameliorate privacy concerns because the warrant’s goal is to eventually connect anonymized data to a person. 110 F.4th at 834 n.9 (5th Cir. 2024). Further, the Fifth Circuit questioned whether this data is truly anonymous when the data can be cross-referenced to reveal users’ identities. Id. In addition, researchers have established that “cross-

26

concern. In this case, police obtained step-two location data for an anonymous device ID. The step-two data revealed that the device user stopped at a gas station. Officers were then able to obtain surveillance footage from the gas station. If the surveillance footage had revealed that only one vehicle with a single occupant pulled up to the gas station, police could have run the plates of the vehicle, found the name and a picture of the registered owner, compared the picture to the person who stepped out of the vehicle on surveillance footage, and thereby positively identified the “anonymous” device ID. The fact that location information can be connected to distinct individuals is of particular concern considering that, as stated earlier, this data has the tendency to reveal the private details about a person’s life.

Given the highly sensitive location information at issue here and the ease with which law enforcement might identify users who were meant to have remained anonymous, we conclude that a person has a reasonable expectation of privacy in location data of the kind stored by Google, whether anonymized or not, under Article I, Section 10, of the Minnesota Constitution. 12

referencing datasets can reveal the identifying information of nearly every ‘anonymized’ user. There are many opportunities to cross-reference an anonymized data dump received from Google, invading the privacy of all users caught up in the geofence.” Amster & Diehl, Against Geofences, 74 Stan. L. Rev. at 409. 12 We do not, however, consider whether there is a reasonable expectation of privacy in any other form of anonymous data.

27

II.

We next turn to the question of whether all geofence warrants are categorically unconstitutional general warrants. Article I, Section 10, of the Minnesota Constitution requires that a search warrant be supported by probable cause and “particularly describ[e] the place to be searched and the person or things to be seized.” Minn. Const. art. I, § 10. The particularity requirement prohibits law enforcement from conducting general or exploratory searches. Sardina-Padilla, 7 N.W.3d at 598. A general warrant “specifie[s] only an offense and le[aves] the decision of whom to arrest and where to search to the discretion of the official executing the warrant.” State v. Jackson, 742 N.W.2d 163, 169 (Minn. 2007). General warrants are problematic because they authorize “a general, exploratory rummaging in a person’s belongings.” City of Golden Valley v. Wiebesick, 899 N.W.2d 152, 162 (Minn. 2017).

Contreras-Sanchez argues that all geofence warrants are categorically prohibited general warrants because they authorize police to search the data of every person whose device registers within a geofence area, regardless of whether there is probable cause to search each person, and because they afford police “unbridled discretion” to target individuals in the geofence in the warrant’s subsequent steps. But if even one geofence warrant could be constitutionally valid, then it cannot be true that all geofence warrants are categorically prohibited. We agree with the court of appeals that it is possible for a geofence warrant to be so narrow that it only captures one person’s data. See Contreras- Sanchez, 5 N.W.3d at 164. The court of appeals provided the following hypothetical as an example of a potentially constitutional geofence warrant:

28

[I]magine that a surveillance camera captures an unidentifiable suspect making a cell-phone call in front of a business that the suspect just burglarized. The phone call occurs in the middle of the night and in a sparsely populated area. The surveillance footage shows no one else in the area for several hours before and after the phone call, and there are no dwellings in the vicinity of where the phone call was made. In this hypothetical situation, law enforcement could obtain a very narrow geofence warrant, limited in both time and location, that virtually guarantees the warrant would only capture the location-history data of the burglary suspect’s cell phone.

Id. Under these circumstances, the geofence warrant would be directed at a single suspect, particularized to that suspect, and capture data only from that suspect. Assuming this hypothetical warrant was also designed to prevent the particularity concerns we discuss below by, for example, requiring independent warrants at each step to limit officer discretion at every stage in the process of gathering and identifying location data within a geofence, then the warrant would be sufficiently particular to remove officer discretion regarding where and whom to search. Thus, because not all geofence warrants are necessarily unconstitutional, we hold that geofence warrants are not categorically unconstitutional general warrants, and courts should assess their validity on a case-by-case basis. 13

13 Only one court—the Fifth Circuit in Smith—has concluded that all geofence warrants are categorically unconstitutional general warrants. 110 F.4th at 837–38. The Fifth Circuit determined that geofence warrants are general warrants because, in step one, law enforcement searches through “troves of location data from hundreds of millions of Google users without any description of the particular suspect” to identify the device IDs present inside the geofence. Id. This argument is unpersuasive. It is Google, not law enforcement, that conducts searches through data to identify the device IDs present inside the geofence. And every search, even a search of physical records, requires a cursory inspection to distinguish records that fall within the scope of the warrant from those that do not. See United States v. Rhine, 652 F. Supp. 3d 38, 82 (D.D.C. 2023) (“[T]he relevant

29

III.

We now turn to the characteristics of the specific warrant in the present case.

Contreras-Sanchez first claims that the warrant lacked probable cause under the United States and Minnesota constitutions. “[N]o warrant shall issue, but upon probable cause . . . .” Minn. Const. art. I, § 10; see also U.S. Const. amend. IV (“[N]o Warrants shall issue, but upon probable cause . . . .”). Probable cause requires a “ ‘fair probability that contraband or evidence of a crime will be found in a particular place.’ ” State v. Wiley, 366 N.W.2d 265, 268 (Minn. 1985) (quoting Illinois v. Gates, 462 U.S. 213, 238 (1983)). Probable cause “does not deal with hard certainties, but with probabilities.” Gates, 462 U.S. at 231. An issuing judge makes a probable cause determination after considering the totality of the circumstances in the search warrant application and accompanying affidavit. State v. Wiggins, 4 N.W.3d 138, 145 (Minn. 2024). In making a probable cause determination, judges may draw on common sense and include reasonable inferences that law enforcement drew from facts. State v. Lester, 874 N.W.2d 768, 771 (Minn. 2016). “We

question is not how Google runs searches on its data, but what the warrant authorizes the Government to search and seize. . . . [Many searches] necessarily . . . require the third party to search some group of records larger than those specifically requested, whether they reside in a file cabinet or on a server.”). Contreras-Sanchez did not argue that the Fifth Circuit’s reasoning on this issue applies in this case. And in any event, the Fifth Circuit affirmed the denial of the motion to suppress evidence derived from the geofence warrant under the good-faith exception. Smith, 110 F.4th at 829, 840.

Here, because we hold below that the warrant violated Contreras-Sanchez’s rights, we do not address the general warrant issue under the Fourth Amendment. If the Supreme Court in Chatrie, however, reaches that issue and decides that a geofence warrant is a general warrant under the Fourth Amendment, we permit the court of appeals to address that issue on remand when addressing the applicability of the good-faith exception and harmless error.

30

review only the warrant application and supporting affidavits to determine if the issuing judge had a substantial basis for concluding that probable cause existed.” Wiggins, 4 N.W.3d at 145 (citations omitted) (internal quotation marks omitted). “When we perform a totality of the circumstances analysis on a warrant application, [w]e defer to the issuing magistrate, recognizing that doubtful or marginal cases should be largely determined by the preference to be accorded to warrants.” Id. at 145–46 (alteration in original) (citations omitted) (internal quotation marks omitted).

This is an issue that we arguably do not need to reach, given that we independently conclude in Part IV that the warrant at issue here fails under the particularity requirement of the Minnesota Constitution. But to the extent it may bear upon the applicability of the good-faith exception on remand, we address the issue of probable cause and hold, rather than simply assume without deciding, that the warrant here was supported by probable cause. In doing so, we are cognizant that the issue of probable cause under the Fourth Amendment is before the Supreme Court in Chatrie. See Chatrie, 136 F.4th 100, cert. granted, No. 25-112, 2026 WL 120676 (U.S. Jan. 16, 2026) (granting review on whether the execution of a geofence warrant violated the Fourth Amendment). Nonetheless, for the sake of completeness—and because the probable cause issue is one that is intrinsically tied to the particular facts of the case, and thus Chatrie’s applicability is uncertain—we address the issue here and conclude that probable cause existed. 14

14 If the Supreme Court in Chatrie decides that the geofence warrant in that case lacked probable cause, we permit the court of appeals to consider on remand what bearing that decision has on our determination that the warrant here was supported by probable cause,

31

Contreras-Sanchez argues, first, that there was no probable cause to support issuing this warrant because the supporting affidavit alleged only that the perpetrators owned cell phones, not that they carried their cell phones at the time Mandujano’s body was moved into the culvert. The warrant application stated that an anonymous informant reported “that other potential suspects . . . have cell phones. [The informant was] unsure what brand of cell phone or who their providers were.” But the issuing judge could have reasonably concluded, given the ubiquity of cell phones, that the perpetrators likely had their cell phones with them at the time Mandujano was killed. Individuals “compulsively carry cell phones with them all the time.” Carpenter, 585 U.S. at 311; see also Riley v. California, 573 U.S. 373, 385 (2014) (“[Cell phones] are now such a pervasive and insistent part of daily life that the proverbial visitor from Mars might conclude they were an important feature of human anatomy.”). Given the pervasiveness of cell phones, and the specific information that the perpetrators of this crime owned them, there was a fair probability that the perpetrators carried their cell phones while transporting the body into the geofence perimeter.

Contreras-Sanchez next argues that there was no probable cause to issue the warrant because the supporting affidavit did not allege that the suspects had used Google’s location tracking technology. The application states, “Your affiant knows Google, LLC retains and uses location information for individuals who use a wide range of Google productions [sic]. This information kept by Google also contains location information, which can be very

and to consider what independent effect such a finding has on the applicability of the good- faith exception to the exclusionary rule.

32

accurate.” While law enforcement’s description of Google’s processes of data collection was sparse, the application needed only to provide a fair probability that evidence of a crime would be found on Google’s servers. See Wiley, 366 N.W.2d at 268. In the present case, the warrant affidavit stated that Google stored location data “for individuals who use a wide range of Google productions [sic].” The issuing judge could have found, based on this language, that there was a fair probability that Google, a large, multinational corporation whose popular applications are used on many types of cell phones, would have collected location data from a perpetrator’s cell phone.

Next, Contreras-Sanchez argues that even if the warrant application established probable cause that evidence of a crime would be found at Google, the application was still insufficient because it was required to demonstrate a probable cause nexus for each person who entered the geofence. Contreras-Sanchez bases his argument on the United States Supreme Court’s decision in Ybarra v. Illinois, 444 U.S. 85 (1979).

In Ybarra, police executed a warrant to search a tavern and a bartender but used the warrant to frisk every patron inside the tavern. Id. at 88. The United States Supreme Court found these searches lacked probable cause because a person’s mere proximity “to others independently suspected of criminal activity does not, without more, give rise to probable cause to search that person.” Id. at 91. The Supreme Court further determined:

Where the standard is probable cause, a search or seizure of a person must be supported by probable cause particularized with respect to that person.

This requirement cannot be undercut or avoided by simply pointing to the fact that coincidentally there exists probable cause to search or seize another or to search the premises where the person may happen to be.

Id.

33

The United States Supreme Court, however, has also long recognized that a lawful search of property may infringe on the privacy interests of an innocent party present at the time of the search. See Zurcher v. Stanford Daily, 436 U.S. 547, 554 (1978). The general rule is that “valid warrants may be issued to search any property, whether or not occupied by a third party, at which there is probable cause to believe that fruits, instrumentalities, or evidence of a crime will be found.” Id. (emphasis in original). “In situations where the State does not seek to seize ‘persons’ but only those ‘things’ which there is probable cause to believe are located on the place to be searched, there is no apparent basis in the language of the [Fourth] Amendment for also imposing the requirements for a valid arrest—probable cause to believe that the third party is implicated in the crime.” Id.

Unlike Ybarra, in which law enforcement searched the patrons themselves, in this case, police sought to search and seize only “things.” Law enforcement established probable cause that evidence of a crime would be found in location data within the confines of the geofence warrant which Google stored; the warrant’s incidental infringement on the privacy interests of other people whose data was captured does not invalidate an otherwise constitutionally sound warrant. The geofence warrant did not need to demonstrate a probable cause nexus for each person who entered the geofence.

The warrant application in this case alleged that a murder had occurred, that an informant told police the perpetrators of the murder owned cell phones, and that Google stored location data for its users across a broad range of products. This established a fair probability that Google’s servers would contain evidence of a crime, thus there was probable cause to issue the geofence warrant.

34

IV.

Finally, we consider Contreras-Sanchez’s claim that the geofence warrant in this case was not sufficiently particular. Because we hold that this requirement was not met under the Minnesota Constitution, we have no need to address whether the warrant was sufficiently particularized under the Fourth Amendment. See Leonard, 943 N.W.2d at 152 n.1; Flowers, 734 N.W.2d at 258. Warrants must “particularly describ[e] the place to be searched and the person or things to be seized.” Minn. Const. art. I, § 10. The Minnesota Constitution, like the Fourth Amendment, “by its terms requires particularity in the warrant, not in the supporting documents.” State v. McNeilly, 6 N.W.3d 161, 176 (Minn. 2024) (emphasis in original). “[The particularity] requirement prohibits law enforcement from engaging in general or exploratory searches.” State v. Bradford, 618 N.W.2d 782, 795 (Minn. 2000). “The particularity requirement . . . prevents the seizure of one thing under a warrant describing another. As to what is to be taken, nothing is left to the discretion of the officer executing the warrant.” McNeilly, 6 N.W.3d at 175 (citation omitted) (internal quotation marks omitted).

“[W]hen determining whether a clause in a search warrant is sufficiently particular, the circumstances of the case must be considered, as well as the nature of the crime under investigation and whether a more precise description is possible under the circumstances.” Id. at 176 (alteration in original). As we recently noted, “[a] warrant permitting a broad search can be sufficiently particular if greater specificity is not possible because officers do not know all the circumstances surrounding the alleged crime.” Sardina-Padilla, 7 N.W.3d at 599. We grant the issuing judge “considerable deference” in determining

35

whether a search warrant is sufficiently particular. See State v. Miller, 666 N.W.2d 703, 713 (Minn. 2003).

We first consider Contreras-Sanchez’s arguments that the geofence warrant was insufficiently particular as to its geographic area and timeframe, and then address his argument that the warrant gave officers “unbridled discretion” in seeking additional data in subsequent steps.

A.

Contreras-Sanchez first argues that the warrant in his case was insufficiently particular because the geofence covered too broad a geographic area. The geofence perimeter encompassed an area 65 feet wide by 290 feet long. It included the culvert where Mandujano’s body was found, a portion of the right of way ditch, and a public road. Contreras-Sanchez contends that the area could have been limited to avoid the public road and to solely include the area around the culvert where Mandujano’s body was found.

Considering the circumstances of the investigation, we conclude that the warrant sufficiently described the place to be searched, which reasonably included a portion of the adjacent road. At the time they requested the warrant, officers had information that multiple people were involved in killing Mandujano. It was reasonable to conclude that one person might have stayed in a vehicle on or near the road to act as a lookout while others moved the victim. Furthermore, the warrant application noted that only a small number of vehicles use the road in question and there were only three residences on that section of the road. Crafting the geofence area to cover a lightly trafficked road was justified given the circumstances of the investigation.

36

Contreras-Sanchez next contends that the warrant was insufficiently particular because its timeframe was too long. The geofence warrant authorized data collection within the geofence for one month, beginning when Mandujano was last seen and ending when his body was found. After being served with the warrant, Google contacted police to say that the month-long timeframe would overload its computers and take considerable time to process. Google suggested providing the data in multiple, smaller timeframes to enable more efficient processing. Law enforcement narrowed the initial search into two shorter time periods per Google’s request. Contreras-Sanchez argues that because law enforcement shortened the initial timeframe into smaller timeframes, it could have been more particular from the outset.

We disagree. On the facts of this case, a longer timeframe satisfied the particularity requirement because officers did not know the specific date Mandujano was killed or moved to the culvert. Officers requested data within a timeframe tailored to the information they had about the crime: the date Mandujano went missing and the date his body was discovered. See Sardina-Padilla, 7 N.W.3d at 601 (concluding that a warrant for approximately three months of defendant’s Facebook activity—including time both before and after the body was discovered—was sufficiently particular because “it was reasonable for investigators to believe that some degree of preparation and communication preceded the crimes” and that content for “a period of time after the crime would yield relevant evidence”). Law enforcement shortened the timeframe not because it had a more particularized suspicion about certain dates, but because it was necessary as a practical

37

means of executing the warrant. 15 The warrant was sufficiently particular with respect to time. 16 B.

Contreras-Sanchez further argues that the warrant was insufficiently particular in its instructions to police because officers were given “unbridled discretion” to choose the device IDs for which to seek additional data in steps two and three. Because we conclude that the warrant’s second step was insufficiently particular, we do not address the question of step three’s particularity.

In step two, the warrant instructed officers to analyze the step-one data for those “who may have witnessed or participated in the [s]ubject [o]ffenses.” It further provided that, “For those accounts identified as relevant to the ongoing investigation through an analysis of provided records, and upon demand, [Google] shall provide additional location history.” Specifically, Google was required to provide two hours of additional location data outside of the geofence, one hour before and one hour after entering the geofence, for

15 Contreras-Sanchez argues that breaking up the month-long timeframe into smaller periods is also evidence that the warrant insufficiently limited officers’ discretion. As previously noted, law enforcement crafted these smaller timeframes—timeframes contained entirely within the month-long timeframe the warrant authorized—in response to the practical realities of obtaining data from Google. There is no requirement that search warrants include “a specification of the precise manner in which they are to be executed.” Dalia v. United States, 441 U.S. 238, 257 (1979). 16 This geofence warrant did authorize collecting data for a longer timeframe than other geofence cases. See, e.g., Info. Stored at Google, 579 F. Supp. 3d at 72 (185 minutes); Rhine, 652 F. Supp. 3d. at 68 (four-and-a-half hours). This does not change the fact that the timeframe was reasonable in Contreras-Sanchez’s case given the circumstances of the investigation.

38

device IDs law enforcement deemed “relevant” with no geographic limits. The warrant application did not indicate law enforcement would seek, and the warrant did not require, an additional warrant before obtaining step-two data.

These instructions did not effectively limit police discretion and authorized impermissible exploratory rummaging in location data. The instructions were deficient in two ways. First, the warrant allowed officers complete discretion to “analyze [the] location data to identify users who may have witnessed or participated in the [s]ubject [o]ffenses.” Second, it left the decision as to which accounts were “relevant” for ongoing investigation entirely to officers’ discretion. Neither determination was subject to judicial review, which would have required a showing of probable cause for a step-two search. Although the officers in this case made a narrow request for data at step two and requested additional data only for a single device ID, the warrant allowed them to request additional data for all 31 device IDs that passed through the geofence. There was nothing in this warrant to prevent law enforcement from treating every device ID in the geofence as belonging to a suspect and rummaging through an additional two hours of location data for all of them. 17 Because there was no requirement to obtain an additional warrant before requesting additional location data at step two, the warrant allowed law enforcement unfettered discretion to interpret step-one data and determine what was relevant without judicial

17 The warrant described the purpose of this additional two hours of location data in step two as “eliminat[ing] outlier points where, from the surrounding data, it becomes clear the reported point(s) are not indicative of the device actually being within the scope of the warrant.” The warrant, however, left it to law enforcement’s discretion to actually limit its use of the data in the way the warrant outlines; it did not prevent law enforcement from accessing more expansive location data without judicial review.

39

review. Accordingly, we hold that this warrant lacked sufficient particularity under the Minnesota Constitution. 18 Because the court of appeals concluded the warrant was sufficiently particular, it did not reach the issue of remedy, including whether the good-faith exception to the exclusionary rule applies in this context and whether the error was harmless. When we reverse a court of appeals decision and are thus confronted with a question the court of appeals did not reach, a remand is typically appropriate. See Dupey v. State, 868 N.W.2d 36, 41 (Minn. 2015); State v. Ross, 732 N.W.2d 274, 276 n.3 (Minn. 2007). We therefore remand to the court of appeals for proceedings consistent with this opinion.

CONCLUSION

For the foregoing reasons, we reverse the decision of the court of appeals and remand to the court of appeals for further proceedings consistent with this opinion.

Reversed and remanded.

GAÏTAS, J., took no part in the consideration or decision of this case.

18 The language in step two was not sufficiently particular and there was no additional warrant requirement before obtaining step-two data. We therefore do not consider whether it is possible to limit police discretion to obtain additional location data with clear instructions in the warrant, or whether an additional warrant is required at each stage of a geofence investigation where law enforcement obtains additional information from a provider storing this type of location data.

C O N C U R R E N C E

HUDSON, Chief Justice (concurring).

I join the majority opinion of the court, except to the extent that the majority suggests that there is no need to stay this case. I write separately to express my view that it would have behooved our court to stay this case pending the Supreme Court’s decision in United States v. Chatrie, 136 F.4th 100 (4th Cir. 2025), cert. granted, No. 25-112, 2026 WL 120676 (U.S. Jan. 16, 2026), which is set for argument on April 27, 2026. Although I question the court’s decision to issue its opinion now, without the benefit of the Supreme Court’s decision in Chatrie, this procedural concern is mitigated in part by our remand to the court of appeals. I emphasize here what the majority acknowledges: that there may be circumstances in which it is appropriate for the court of appeals on remand to consider the Supreme Court’s decision in Chatrie.

C-1

DISSENT

MCKEIG, Justice (dissenting).

The issue of whether geofence warrants violate the Fourth Amendment is pending before the United States Supreme Court in Chatrie v. United States, 136 F.4th 100 (4th Cir. 2025), cert. granted, No. 25-112, 2026 WL 120676 (U.S. Jan. 16, 2026). Following an en banc hearing, the members of the United States Court of Appeals for the Fourth Circuit adopted different views of whether the reasoning in Carpenter v. United States, 585 U.S. 296 (2018), applied with equal force to geofence warrants. Contrast Chatrie, 136 F.4th at 138–41 (Richardson, J., concurring) (concluding that, unlike the government’s actions in Carpenter, the execution of the geofence warrant did not constitute a search under the Fourth Amendment), with Chatrie, 136 F.4th at 115, 120–30 (Wynn, J., concurring) (concluding that, like the government’s actions in Carpenter, the execution of the geofence warrant constituted a search under the Fourth Amendment), and Chatrie, 136 F.4th at 144, 148–53 (Berner, J., concurring) (concluding that, under Carpenter, the execution of the geofence warrant constituted a search under the Fourth Amendment when the government obtained non-anonymous location data but not when the government obtained anonymous information unlikely to be traced to a specific individual), and Chatrie, 136 F.4th at 141– 43 (declining to determine whether there was a Fourth Amendment violation and instead concluding that the exclusionary rule did not apply because the constitutionality of geofence warrants was unresolved when the warrant was obtained and the purpose of the

D-1

exclusionary rule was not met). Unable to resolve those differences, the members of the Chatrie court never formed a majority opinion. 1 The Supreme Court is scheduled to hear argument in Chatrie later this month, on April 27, 2026. In the past, we have cautiously taken “a more restrained approach when both constitutions use identical or substantially similar language” and have departed from this restrained approach only after the United States Supreme Court has provided guidance. City of Golden Valley v. Wiebesick, 899 N.W.2d 152, 157 (Minn. 2017) (providing three situations where we will depart from the Supreme Court’s guidance when interpreting substantially similar language); see also Kahn v. Griffin, 701 N.W.2d 815, 824–25 (Minn. 2005) (discussing our approach when interpreting the Minnesota Constitution). And we generally “do not construe our state constitution as providing more protection for individual rights than does the federal constitution unless there is a principled basis to do so.” State v. Ezeka, 946 N.W.2d 393, 402 (Minn. 2020) (citation omitted) (internal quotation marks omitted).

The court claims there is no need wait for the Supreme Court’s impending guidance because its analysis rests on an adequate and independent state ground and because the court has a responsibility to decide the issues presented without further delay, not only for

1 The Fourth Circuit used the word “fractured,” Chatrie, 136 F.4th at 157 (Gregory, J., dissenting), to describe its reasoning—nine opinions (eight concurrences and one dissent) accompanied a one-sentence per curiam order in which 14 of 15 judges voted to affirm the district court’s decision not to exclude the geofence evidence, see Chatrie, 136 F.4th 100. The Chief Judge described this as a “labyrinth” of opinions, representing “widely divergent views on the intersection of the Fourth Amendment and the groundbreaking investigative tool at issue here.” Chatrie, 136 F.4th at 108, 101 (Diaz, C.J., concurring).

D-2

the parties but for the people of Minnesota. The court’s claim does not withstand scrutiny for three reasons. First, the court’s discussion of the existing law on electronic data relies entirely on federal cases, including Carpenter, 585 U.S. 296. 2 Second, appellant Ivan Contreras-Sanchez has informed the court that he does not object to staying this case pending the Supreme Court’s decision in Chatrie. Third, the Minnesota Legislature is now considering the competing public policy concerns of a proposed statute that would prohibit government entities from requesting or obtaining reverse-location information, similar to the information here, in non-emergency situations. S.F. 1120, 94th Minn. Leg. 2026 (as introduced); H.F. 103, 94th Minn. Leg. 2026 (as introduced).

I acknowledge that, because we are evenly divided on whether we should wait for the Supreme Court’s impending guidance, no stay may issue and the court has the power to move forward in this case. But in my view, it is highly imprudent to exercise that power and move forward without the benefit of the impending guidance from the Supreme Court regarding Carpenter, a case that divided the judges in Chatrie, see, e.g., 136 F.4th at 101

2 The concurrence agrees that the more prudent approach in this case would be to stay further proceedings pending the Supreme Court’s decision in Chatrie. The concurrence questions the court’s decision to “issue its opinion now, without the benefit of the Supreme Court’s decision in Chatrie.” Nevertheless, it concludes that a stay is not required here in part because the court is remanding to the court of appeals. The court’s remand directs the court of appeals to consider the State’s argument regarding the good-faith exception to the warrant requirement and only allows the court of appeals to revisit this court’s probable cause analysis if the Supreme Court in Chatrie decides that the geofence warrant in that case lacked probable cause. To be clear, if the Supreme Court in Chatrie decides that the reasoning of Carpenter, which the court relies upon to reach its decision in Part I, does not apply to a geofence warrant, nothing in the remand order or this court’s opinion will allow the court of appeals to revisit that important issue—an issue that fractured the judges in Chatrie.

D-3

(Diaz, C.J., concurring), and which this court cites 30 times in its opinion, especially when (1) the court has failed to establish an adequate and independent state ground for its decision, (2) Contreras-Sanchez does not object to a stay, and (3) the Legislature is now weighing the relevant public policies. Consequently, I dissent.

A.

In reversing Contreras-Sanchez’s conviction, the court claims it is interpreting Article I, Section 10, of the Minnesota Constitution independently of the United States Supreme Court’s articulation of Fourth Amendment protections. I disagree.

The court’s claim that its interpretation of the Minnesota Constitution rests solely on existing state case law is belied by the court’s in-depth discussion of Carpenter and Carpenter’s omnipresence throughout the court’s discussion of electronic data in Part I. The court’s discussion of the existing law on electronic data relies entirely on federal cases, including Carpenter, which the court mentions 30 times in its opinion. Moreover, in Pennsylvania v. Labron, 518 U.S. 938, 940–41 (1996), the United States Supreme Court concluded that citations to state cases that are grounded in federal precedent do not establish an adequate independent state ground when it is not clear that the federal precedent was used for guidance alone, rather than compelling the result. All of this court’s discussion of a person’s expectation of privacy in electronic data is grounded in Carpenter. It is true that the court cites In re B.H., 946 N.W.2d 860, 869 (Minn. 2020), for the proposition that cell phones can track a person’s location history with alarming accuracy, revealing not only the details of a person’s movements, “but through them his familial, political, professional, religious, and sexual associations.” The parenthetical that follows

D-4

the court’s citation, however, omits the source of the quoted language—Carpenter, 585 U.S. at 311. The court’s citation to State v. Sardina-Padilla, 7 N.W.3d 585, 599 (Minn. 2024), also fails to establish an adequate independent state ground because our decision in Sardina-Padilla cites in support only B.H., 946 N.W.2d at 869, with a parenthetical quoting language from Carpenter.

The court also explicitly relies on Carpenter in its assessment of the third-party doctrine, arguing that the approach in this case and in State v. Leonard, 943 N.W.2d 149, 155 (Minn. 2020), is “similar.”

As the above examples demonstrate, the court’s entire analysis inextricably relies on the reasoning in Carpenter; therefore, the court has failed to establish an adequate and independent state ground for its decision. This alone should counsel against moving forward without the benefit of the impending guidance from the Supreme Court in Chatrie—a case in which the judges adopted different views of whether the reasoning in Carpenter applied with equal force to geofence warrants. See, e.g., Chatrie, 136 F.4th at 108, 101 (Diaz, C.J., concurring). When one additionally considers that Contreras-Sanchez does not object to a stay and that the Legislature is now weighing the relevant public policies, it is highly imprudent to move forward without the benefit of the impending guidance from the Supreme Court.

B.

For all the reasons stated above, I decline to reach the merits of the issue that the court chooses to resolve now. Without deciding the issue, however, I further observe that

D-5

the facts of this case and the decisions by the courts below reinforce that the court’s decision here is not a foregone conclusion.

Manuel Mandujano went missing on March 25, 2021. About four weeks later, law enforcement officers found Mandujano’s brutally beaten body in a drainage culvert in rural Dakota County. His hands were bound behind his back and a nail was embedded in the heel of one of his feet. An anonymous informant told police they knew someone involved in the murder of Mandujano, and that multiple individuals participated in the act. The anonymous informant said that the suspects involved in the murder had cell phones, though they were unsure which brand of cell phone or which providers the participants used.

On April 29, 2021, law enforcement applied for a geofence warrant to obtain anonymized location data stored in Google’s internal database housing users’ location history data (“Google’s internal database”). 3 To be clear, the application did not seek a warrant to search the contents of Contreras-Sanchez’s cell phone. Instead, the application sought a warrant to conduct an anonymized search of location data collected from a rural highway ditch and stored in Google’s internal database where Google stores the location data that it receives from the cell phones of people who have authorized Google to collect that data by affirmatively opting in to its collection. 4 The requested geofence warrant

3 “Anonymized location data” in this context refers to geographic information derived from a device or user that has been processed to remove personally identifiable details so that the data cannot reasonably be used to identify a specific device or user. 4 The warrant application requested location data “[i]ncluding, but not limited to[,]

GPS, WiFi or Bluetooth, and/or cell tower sourced location history data.” The geographic parameters specified in the warrant application “encompass[ed] a public roadway and a

D-6

sought anonymized location data collected from a 65-foot by 290-foot rectangular perimeter in southern Dakota County that encompassed the culvert where Mandujano’s body was located, a portion of the public road, and a portion of the right-of-way ditch between March 25 (the day Mandujano went missing) and April 26 (the day his body was found). 5 The following photograph depicts the rural highway ditch and surrounding area:

portion of a right of way ditch area.” And the warrant application acknowledged that the location data requested would only be from “individuals who use a wide range of Google productions [sic].” The warrant application, therefore, acknowledged that any data returned in response would only be data generated by individuals who had affirmatively opted into Google’s policies by using Google’s products. 5 In the warrant application, law enforcement asserted that a small number of vehicles used the public road in question at any given time and there were only 3 residences on that section of roadway, with the closest located over 1,200 feet away.

At the pretrial evidentiary hearing where Contreras-Sanchez challenged the geofence warrant, a detective described the remoteness of the location, stating that the road on which Mandujano’s body was found was seldom traveled. The detective also explained that Mandujano’s body was found only a half-mile from where the paved road terminated at a gravel road. The detective gave similar testimony at trial.

D-7

The requested geofence warrant also limited the release of this narrowly focused location data using a three-step process. After reviewing the application, a district court issued the requested warrant.

After Google received the geofence warrant, it complied with step one of the authorized process, which required Google to provide the investigating officers anonymous user identification numbers (device IDs) for every electronic device that reported its presence within the geofence perimeter between the date Mandujano went missing and the

D-8

date his body was found. 6 The anonymized step-one data revealed that during the seven- day period between March 25, 2021, and March 31, 2021, only 12 electronic devices reported their location in the geofence perimeter. 7 Eleven of the devices only reported their presence in the geofence perimeter once. The officers took no further action regarding those devices. The 12th device, however, reported its location in the geofence perimeter 46 times during a 10-minute period on March 29, 2021. Having reasonably determined that the 12th device was relevant to the ongoing investigation, the officers proceeded to step two of the process authorized under the geofence warrant.

In accordance with step two, Google provided the investigating officers with additional anonymized location data from Google’s internal database for the 12th device. More specifically, Google disclosed all the anonymized location data from the 12th device for the 60 minutes before and after the device reported its location within the geofence perimeter. The purpose of this step was to eliminate outlier data points where it became clear the device was not within the scope of the warrant. The anonymized step-two location data reported by the 12th device revealed that the device was present at a gas station 40

6 In its initial response to the investigating officers, Google noted that the information did not include any information regarding cell-sourced location coordinates because the specified search area was too small for such information to be accurate. 7 Although the timeframe specified in the search warrant was from the date that Mandujano’s family last saw him (March 25, 2021) until the date that Mandujano’s body was found (April 26, 2021), Google contacted the warrant applicant and asked him to narrow the timespan so Google could process the data faster. In response to Google’s request, the warrant applicant narrowed the timeframe down to two seven-day periods: March 25–31, 2021, and April 1–7, 2021. Because all the relevant evidence was obtained during the first seven-day period, the second seven-day period is not relevant here.

D-9

minutes before it reported its presence in the geofence perimeter. The gas station provided surveillance video to the investigating officers, which showed four persons of interest at the gas station. An investigating officer recognized one individual and one vehicle from the ongoing investigation. Based on the anonymized step-two data, the investigating officers moved to the third step in the process authorized by the geofence warrant.

Although the third step in the process authorized by the geofence warrant already required Google to provide the investigating officers information identifying the owner of the 12th device without obtaining another warrant, the investigating officers here nevertheless applied for a second warrant. This second warrant directed Google to provide the identifying information based on the information obtained during the officers’ investigation, including the anonymized first- and second-step location data. After reviewing the application, a district court issued the second warrant. When Google received the second warrant, Google identified the 12th device as belonging to Contreras- Sanchez.

Police spoke with Contreras-Sanchez several times before ultimately arresting him in November 2021. In an interview with police, Contreras-Sanchez admitted participating in the murder of Mandujano and showed police videos on his phone depicting a badly beaten Mandujano. In one video, Contreras-Sanchez asked the victim in Spanish, “that’s what you get for being a snitch, right?”

The State charged Contreras-Sanchez with second-degree intentional murder and second-degree unintentional felony murder. Contreras-Sanchez pleaded not guilty and demanded a jury trial.

D-10

Before trial, Contreras-Sanchez moved to suppress the 10 minutes of location data collected from Google’s internal database that showed his presence in the geofence perimeter. Contreras-Sanchez argued that all geofence warrants are per se unconstitutional general warrants and, in the alternative, the geofence warrant in his case lacked probable cause, was not sufficiently particular, and was overbroad. The State argued that Contreras- Sanchez failed to establish that he had a reasonable expectation of privacy in the ten minutes of anonymized location data collected from the geofence perimeter and stored in Google’s internal database. In the alternative, the State argued that the geofence warrant here was supported by probable cause and was sufficiently particular.

On May 17, 2022, the district court held a contested suppression hearing. The hearing provided Contreras-Sanchez an opportunity to present evidence to satisfy his burden to prove that he exhibited an actual subjective expectation of privacy in the location data collected from the geofence perimeter, as required to prove there was a search under the Fourth Amendment or Article I, Section 10. Kyllo v. United States, 533 U.S. 27, 33 (2001) (“[A] Fourth Amendment search occurs when the government violates a subjective expectation of privacy that society recognizes as reasonable.”); Leonard, 943 N.W.2d at 156 (“Under Article I, Section 10, a search occurs when law enforcement intrudes upon an individual’s subjective expectation of privacy that society is prepared to recognize as reasonable.”). Contreras-Sanchez chose not to testify at the hearing. He failed to present testimony from any Google employee on whether the affirmative opt-in choice to location information storage was confusing or illusory. He also failed to provide any testimony whatsoever from a Google employee, law enforcement, or an expert witness regarding data

D-11

privacy or technology. In fact, Contreras-Sanchez did not call any witnesses or submit any documents at the suppression hearing. The State presented testimony from four law enforcement officers who described the investigation, the process of obtaining the geofence warrant, and the process of analyzing the data obtained through the geofence warrant. Unsurprisingly, none of the law enforcement officers testified about Contreras-Sanchez’s subjective expectation of privacy in the ten minutes of location data collected from the geofence perimeter and stored in Google’s internal database, or about whether the affirmative opt-in choice was confusing or illusory. After considering the evidence presented at the suppression hearing, the district court denied Contreras-Sanchez’s motion to suppress. Following a jury trial, Contreras-Sanchez was convicted of second-degree intentional murder and the district court imposed a 480-month sentence.

On appeal, Contreras-Sanchez argued that the district court abused its discretion when it denied his motion to suppress. The court of appeals concluded that even if it assumed “without deciding that Contreras-Sanchez had a reasonable expectation of privacy in his anonymous location-history data,” he was not entitled to any relief because “the geofence warrant passe[d] constitutional muster with respect to its authorization of the seizure of anonymous location-history information.” State v. Contreras-Sanchez, 5 N.W.3d 151, 163, 162 (Minn. App. 2024).

Particularly against this factual record and the holdings by the courts below, it would have been prudent for the court to wait for the Supreme Court’s decision in Chatrie before reversing both the court of appeals and the district court.

D-12

* * *

In sum, I am deeply concerned by the court’s decision to issue our opinion now, without waiting for the impending guidance of the Supreme Court in Chatrie, which is scheduled for oral argument later this month. I acknowledge that, because we are evenly divided on whether we should wait for the Supreme Court’s guidance, no stay may issue and the court has the power to move forward here. But in my view, it is highly imprudent to move forward without the benefit of the impending guidance from the United States Supreme Court regarding Carpenter, a case that divided the judges in Chatrie and which this court cites 30 times in its opinion, especially when (1) the court has failed to establish an adequate and independent state ground for its decision, (2) Contreras-Sanchez does not object to a stay, and (3) the Legislature is now weighing the relevant public policies. This is particularly true where—as illustrated by the facts of this case and the decisions of the courts below—the court’s decision here is not a foregone conclusion.

For the reasons stated above, I dissent.

MOORE, III, Justice (dissenting).

I join in the dissent of Justice McKeig.

D-13