Location records behind a warrant: the 2018 cell-site ruling made legal process a data classification job

On June 22, 2018, the Supreme Court required a warrant for historical cell-site records and left other location-revealing records open. Carriers, banks, health systems and agencies answering demands had to classify their own location data, and a 2026 ruling extended the reasoning.

Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.

Two-minute summaryThe dek and the first sentence of each section, quoted from the article

On June 22, 2018, the Supreme Court required a warrant for historical cell-site records and left other location-revealing records open. Carriers, banks, health systems and agencies answering demands had to classify their own location data, and a 2026 ruling extended the reasoning.

  1. The week On Friday, June 22, 2018, the Supreme Court held that the government’s acquisition of historical cell-site location records from wireless carriers was a search under the Fourth Amendment.
  2. What moved: detailed location over time The first part concerns the category the Court created.
  3. What stayed: the records and techniques left alone The second part concerns what the Court said it was not deciding.
  4. The holders: classifying the record before answering the demand The third part concerns organizations that receive demands.
  5. The requesters: what agencies had to change The fourth part concerns agencies that make demands.
  6. Objections, and where the argument is weakest Four objections test the argument that the decision made legal process a classification problem.
  7. Checks for a records custodian and an investigating agency The four parts translate into checks that apply on both sides of a demand.
  8. With hindsight Questions the Court left open in 2018 returned through geofence warrants.

The week

On Friday, June 22, 2018, the Supreme Court held that the government’s acquisition of historical cell-site location records from wireless carriers was a search under the Fourth Amendment. Before compelling a carrier to turn over those records, the Court said, the government’s obligation is a familiar one: get a warrant. The case was No. 16-402, on review from the Sixth Circuit, and the vote was five to four.

The facts began with a robbery investigation. After the FBI identified the cell phone numbers of several suspects, prosecutors obtained court orders under the Stored Communications Act directing wireless carriers to disclose cell-site records. Under that statute, a court could order disclosure when the government offered “specific and articulable facts” showing reasonable grounds to believe the records were relevant and material to an ongoing criminal investigation. One carrier produced records spanning 127 days. In all, the government obtained 12,898 location points for one defendant’s phone, an average of 101 a day.

The defendant moved to suppress the records, arguing that obtaining them without a warrant supported by probable cause violated the Fourth Amendment. The district court denied the motion and the records were used at trial. The Sixth Circuit affirmed, reasoning that the defendant lacked a reasonable expectation of privacy in location information he had shared with his carriers.

Chief Justice Roberts wrote for the Court, joined by Justices Ginsburg, Breyer, Sotomayor and Kagan. The opinion said the records sat at the intersection of two lines of cases: one on a person’s expectation of privacy in his physical movements, and one on information voluntarily turned over to third parties. The Court declined to extend its third-party precedents, which concerned bank records and dialed telephone numbers, to cell-site records. It described those records as a detailed, encyclopedic and effortlessly compiled chronicle of a person’s movements.

The Court held that accessing seven days of such records was a search, and said it need not decide whether some shorter period could be obtained without one. It said an order under the statute’s “specific and articulable facts” standard was not a permissible mechanism for obtaining historical cell-site records. It added that case-specific exceptions, such as exigent circumstances, could still support warrantless collection, noting that lower courts had approved warrantless searches related to bomb threats, active shootings and child abductions.

The Court called its decision a narrow one. It expressed no view on real-time cell-site data or “tower dumps,” downloads of information on all devices that connected to a cell site during an interval. It said it did not disturb its bank-record and dialed-number precedents, or call into question conventional techniques such as security cameras. It did not address other business records that might incidentally reveal location, or collection involving foreign affairs or national security.

Four justices dissented in four separate opinions. Justice Kennedy, joined by Justices Thomas and Alito, warned that the decision would limit an important investigative tool. He also wrote that courts and officers would have to guess how much information could be requested before a warrant was required. Justices Thomas, Alito and Gorsuch also wrote dissents.

The June 2018 cell-site decision was a ruling about what the government must do. Its working parts landed elsewhere. The records were held by carriers. The demands for them arrived at carriers’ legal process units. After the decision, the question of what process a demand required depended on what kind of record was being sought. That made every holder of location data, and every agency asking for it, part of the decision’s machinery.

Our argument is that the decision turned legal process into a data classification problem. Before it, a holder of business records could answer a law enforcement demand for those records by checking that the process was facially valid. After it, a holder had to know whether the record requested fell into the category the Court had placed behind a warrant, the categories it had left alone, or the open space between. Banks, health systems, carriers and app providers all hold data that could fall in each.

The analysis has four parts. The first is what moved: the category of records the Court placed behind a warrant, and why. The second is what stayed: the records and techniques the Court expressly left under existing law. The third is the holders: what the decision asked of organizations that receive demands for records. The fourth is the requesters: what it asked of the agencies that make those demands. A section on objections and a closing section of checks follow.

The parts do not overlap. The first two describe the legal line. The third describes the work on the receiving side of a demand, and the fourth the work on the sending side. A bank can be on the receiving side; a state agency can be on both, as a holder of records and as an investigator. Nothing here is legal advice. Whether a particular demand requires a warrant is a question for counsel on the facts.

The sector label here is cross-sector because the same line runs through financial services, health care and government. The lens is data because the decision’s effect inside an organization is a question about how data is classified, retained and released.

What moved: detailed location over time

The first part concerns the category the Court created. It is narrower than “location data” and broader than one carrier’s records.

The reasoning

The Court began from two lines of precedent. One recognized that people have a reasonable expectation of privacy in the whole of their physical movements. The other held that people have no reasonable expectation of privacy in information they voluntarily convey to third parties, such as bank records or the numbers they dial. Cell-site records belonged to both: they tracked movement, and they were held by a business.

The Court resolved the conflict by looking at the nature of the records. It said cell-site records hold for many Americans the “privacies of life,” and give the government near perfect surveillance that can travel back in time, limited only by carriers’ retention policies. It contrasted that with the limited types of personal information in its earlier third-party cases. It also said the information is not truly shared. Carrying a cell phone is indispensable to participation in modern society, and the phone logs a cell-site record by its operation, without any affirmative act by the user beyond powering up.

The features that defined the category

The opinion’s reasoning turned on features of the data rather than its label. The records were detailed. They were comprehensive, covering a person’s movements over days or months. They were compiled effortlessly, as a by-product of the service. They were retrospective, letting the government reconstruct the past. And their collection was not, in a meaningful sense, voluntary.

Those features describe more than cell-site records. They describe any data set that logs a person’s location continuously, for a long period, as a side effect of using an ordinary service. In our reading, that is why the decision mattered beyond carriers. A record holder deciding whether a demand falls within the decision’s logic can test its data against those features.

Why seven days did not settle the line

The parties had offered the Court bright lines. In a footnote, the Court said one side proposed a 24-hour cutoff and the government suggested seven days. The government treated the seven days requested from one carrier as the relevant period, though that carrier produced only two days of records. The Court held that accessing seven days of records was a search and declined to say whether any shorter period was free of Fourth Amendment scrutiny.

For record holders, that footnote mattered. A demand for three days of records, or one day, was not clearly inside or outside the holding. Justice Kennedy’s dissent read the decision as requiring a warrant for more than six days of records and criticized the Court for not explaining why shorter periods might differ. A legal process team could not resolve that debate. It could only adopt a policy, with counsel, on how it would treat short-period demands.

The same features in other sectors’ data

Applied outside carriers, the features sort data sets quickly. A bank’s card transaction history reveals location only at the moments of purchase, and the customer chose each transaction. A mobile banking app that collects device location at every session, or continuously for fraud scoring, generates a more frequent log as a by-product of using the service. A hospital’s wireless network may log which access point a patient’s phone connected to over a stay. A badge system logs entries by door and time.

None of those records was before the Court, and the Court said it did not address business records that incidentally reveal location. But the features let a holder rank its data sets from least to most like the records the Court protected. That ranking is the starting point for a classification guide, and for deciding which data sets need counsel’s attention first.

The process the category requires

The Court held that an order under section 2703(d) of the Stored Communications Act, issued on “specific and articulable facts,” was not enough. The government had described the gap between that standard and probable cause as “gigantic,” and the Court used the word. For historical cell-site records, the government needed a warrant supported by probable cause, subject to case-specific exceptions such as exigency.

The statute itself already offered both routes. Section 2703(c) lets a government entity require a provider to disclose customer records if it obtains a warrant, or a court order under subsection (d), among other routes. The decision did not rewrite the statute. It held that for this category, only one of the statute’s routes satisfied the Constitution.

What stayed: the records and techniques left alone

The second part concerns what the Court said it was not deciding. The list is long, and for record holders it is as important as the holding.

The bank-record and dialed-number precedents

The Court said it did not disturb the application of its earlier third-party cases. One of those had held that a depositor had no Fourth Amendment claim in his bank’s records of his checks, deposit slips and statements, which the Court had described as business records of the bank. The other concerned the numbers dialed from a telephone.

For banks, that meant the constitutional baseline for ordinary account records stayed where it was. The statutory baseline also stayed. Congress enacted the Right to Financial Privacy Act in 1978. It bars federal agencies from obtaining a customer’s financial records from a financial institution unless the records are reasonably described and the access comes through one of several routes. The routes include customer authorization, an administrative subpoena or summons, a search warrant, a judicial subpoena or a formal written request, each meeting the statute’s requirements.

The Act defines “Government authority” as an agency or department of the United States, or its officers, employees or agents. In our reading, state and local demands therefore turn on state law and the Fourth Amendment rather than the Act. A bank’s legal process team therefore already worked with different rules for different requesters. The cell-site decision added a question about the record itself.

Subpoenas in the overwhelming majority of cases

The Court went out of its way to reassure. It said the government would be able to use subpoenas to acquire records in the overwhelming majority of investigations. A warrant was required, it said, only in the rare case where the suspect has a legitimate privacy interest in records held by a third party.

That sentence sets the default for holders. The bulk of demands, in the Court’s own expectation, would still be answerable on a subpoena. The work the decision created lay at the margin: recognizing the rare demand that seeks a record of the protected kind. A holder that treated every demand as requiring a warrant would misread the decision as surely as one that treated none that way.

Real-time data and tower dumps

The Court expressly left open real-time cell-site data and tower dumps. A tower dump returns information on every device that connected to a site during an interval. It is broad across people and narrow in time, the opposite profile from the records at issue. The Court did not say whether it required a warrant.

For carriers, those open questions meant that some demands after June 2018 sat in uncertain territory. A carrier could comply with the process presented, challenge it or seek clarification. Each choice had costs. The decision gave no rule for those demands, only reasoning that might or might not extend to them.

Conventional techniques and incidental location

The Court also said it did not call into question conventional surveillance techniques and tools, such as security cameras. And it did not address other business records that might incidentally reveal location information. That second phrase covers a large share of the records institutions hold.

A card transaction records where a purchase was made. An automated teller machine withdrawal records a location. A hospital admission records where a patient was. A building access log records who entered which door. Each reveals location, but as a by-product of a transaction rather than as a continuous log. The Court left those records under its existing law, at least for the time being.

Exigency and national security

Finally, the Court preserved exceptions. Exigent circumstances could justify warrantless collection of cell-site records, and the Court mentioned bomb threats, active shootings and child abductions as examples lower courts had accepted. The opinion also did not consider collection techniques involving foreign affairs or national security. Each exception is a path by which a demand without a warrant could still be lawful, and each requires the holder to judge whether the path is being used properly.

The holders: classifying the record before answering the demand

The third part concerns organizations that receive demands. The decision did not regulate them directly. It changed what their legal process teams had to decide.

Carriers

Carriers were the immediate recipients. After June 22, 2018, a demand for historical cell-site records supported only by a section 2703(d) order fell short of what the Court required. A carrier’s legal process unit needed to recognize those demands and respond accordingly, and to apply the open questions on real-time data and tower dumps consistently.

That work is classification. Each demand has to be matched to the type of record it seeks, the period it covers and the process that supports it. Carriers that had built their intake around the statute’s categories now needed a category defined by the Court’s reasoning.

The decision also affected demands already in progress. Records sought under orders issued before June 22, 2018, and not yet produced, raised the question whether to produce on the existing order or wait for a warrant. That question was for counsel and the requesting agency. The point for operations was that intake had to recognize such demands and hold them for a decision.

Banks

Banks hold location-revealing records of two kinds. Transaction records, such as card purchases and teller machine withdrawals, reveal location incidentally. The Court left records like those under existing law. Digital banking can create a different kind: device location or network data collected through mobile apps or online sessions, for example to detect fraud. Depending on how it is collected and kept, such data could resemble the continuous, by-product logs the Court described.

A bank’s legal process team therefore faces a classification step that its account records never required. When a demand seeks app or device data, the team should know what the bank collects, how often, for how long it keeps it and where it lives. Only then can counsel judge whether the demand seeks a record of the kind the Court placed behind a warrant. For federal demands, the Right to Financial Privacy Act still applies to financial records, whatever the constitutional analysis.

Broker-dealers and wealth managers inside banking groups hold similar data through trading apps and client portals. Their legal process teams may receive requests from securities regulators as well as law enforcement. The classification step is the same: identify which data sets record location continuously as a by-product of service, and route demands for those to counsel.

Health systems

Health systems face the question under HIPAA. The Privacy Rule permits a covered entity to disclose protected health information for law enforcement purposes in defined circumstances. Under 45 CFR 164.512(f), those include compliance with a court order or court-ordered warrant, a subpoena or summons issued by a judicial officer, or a grand jury subpoena. They also include an administrative request, if the information is relevant and material, the request is specific and limited, and de-identified information could not reasonably be used.

The HIPAA rule permits disclosures; it does not decide whether the government needed a warrant to make the demand. A health system that holds location-rich data, such as readings from remote monitoring devices, location data from a patient app or detailed access logs, has two questions to answer. One is whether HIPAA permits the disclosure. The other is whether the demand is constitutionally sufficient for the kind of record it seeks. The decision made the second question live for some records.

App providers and vendors

Organizations can also hold location data through vendors. A bank’s fraud vendor, a health system’s patient engagement platform or an agency’s mobile app provider may hold the location records, not the institution. A demand may go to the vendor directly.

Institutions should know what their vendors would do. Contracts can require notice to the institution when a vendor receives a demand for the institution’s data, where the law permits notice. They can require the vendor to refer demands to the institution, or to apply an agreed classification before responding. Without such terms, the institution may learn of a disclosure only afterward.

Preservation before process

Demands do not always begin with process. Section 2703(f) of the Stored Communications Act requires a provider of electronic communication or remote computing services, on a government request, to preserve records pending a court order or other process. The statute sets a 90-day retention period, extendable for another 90 days on renewed request.

Preservation separates two decisions that classification has to keep apart. Preserving records does not disclose them. A holder can preserve on request while it waits for the process its classification requires. For holders covered by the statute, that sequence lets the organization protect evidence without disclosing a protected record on insufficient process. Holders outside the statute should ask counsel how they treat preservation requests.

Classification needs a process to live in. A legal process function, in our view of good design, has five elements. It has a single intake point for demands, so none arrive unseen at a branch, a clinic or a vendor. It logs each demand with the requester, the legal authority cited, the records sought and the period covered. It classifies the records sought against the organization’s inventory. It escalates to counsel any demand that seeks a protected or uncertain category. And it records what was produced, when and on what basis.

The log is the evidence. If a customer, patient, regulator or court later asks why a record was disclosed, the organization can show the demand, the classification and the decision. If a demand was refused or narrowed, the log shows why. Without the log, an organization’s account of its handling depends on memory.

How the burden differs by organization

A large bank or national carrier may receive enough demands to justify a dedicated team and tooling. A credit union or community bank may receive few, handled by a compliance officer with counsel on call. A broker-dealer can receive demands from regulators as well as law enforcement, under different authorities. A health system can receive demands at hospitals, clinics and its privacy office, and must route them under HIPAA’s rules.

A county or city can receive demands for its own records, from police in its own jurisdiction and from other agencies. For a small government, the risk is that demands are answered by whichever department holds the data, without central review. The classification step requires that someone with authority sees each demand for location-rich records before it is answered.

What classification requires inside an organization

Across sectors, the decision asked the same thing of holders: know your location data. That means an inventory of data sets that record location, with the features the Court cared about. For each data set, the inventory records how detailed it is, how continuous and how long it is kept. It also records whether the data is generated as a by-product of service or by a deliberate act of the customer.

The inventory should cover systems as well as data sets. Location data can sit in production databases, analytics copies, backups and logs. A demand for one customer’s history may be answerable from a log the business team does not know it keeps. The inventory is complete only when it includes those copies.

The inventory serves two purposes. It lets the legal process team classify demands quickly and consistently. And it lets the organization decide whether to keep such data at all. Data that is not retained cannot be demanded. Data minimization, argued on privacy and security grounds, gains a legal process argument as well.

The requesters: what agencies had to change

The fourth part concerns agencies that make demands. Federal, state and local law enforcement, inspectors general and regulators with investigative powers all request records from third parties.

Warrants instead of orders

For historical cell-site records, the change was direct. An investigator who would have sought a section 2703(d) order now needed a warrant on probable cause, or a documented exception. That moved the request earlier or later in an investigation, depending on when probable cause existed. It also changed the paperwork, the reviewing officer and the record the agency kept.

Agencies with case management systems had to reflect the change. A request template that offered a 2703(d) order for cell-site records needed revision. Training had to explain why the same carrier records now required different process. Where agencies relied on exigency, they needed to record the facts that supported it at the time, because a court would later judge them.

A warrant also has to be drafted for data. Probable cause has to connect the records sought to the offense, and the warrant has to describe them with particularity: which account or device, which period, which data fields. Broad windows invite later challenge. An agency that drafts warrants for location data narrowly, and explains why each window and field is needed, gives the reviewing magistrate and any later court a clearer record.

Regulators and inspectors general

Not every requester is a police department. Financial regulators, inspectors general and other investigative bodies use administrative subpoenas and other process to obtain records. For federal agencies seeking financial records, the Right to Financial Privacy Act sets the routes, including administrative subpoenas and summonses that meet its requirements.

The 2018 decision did not address those authorities directly, and its discussion concerned criminal investigation. Whether its reasoning reaches a particular administrative demand for location-rich data is a question for the agency’s counsel. An agency that wants its demands to hold up should ask the question before issuing them rather than after a recipient objects.

Data the agency itself holds

Agencies are also holders. A state department of motor vehicles, a transit authority with fare card records, a toll authority, or a public health agency running an app may hold location records about residents. Requests from other agencies for those records raise the same classification questions a bank faces. The fact that both parties are public bodies does not settle what process is needed.

That makes the decision an internal governance matter for state and local government. An agency that receives requests from police for location-rich records should have a written policy on what process it requires, reviewed by counsel in light of the decision and later cases.

Data obtained without compulsion

The decision concerned records the government compelled a carrier to disclose. It did not address location data an agency obtains by other means, such as by purchase from a commercial source or by voluntary disclosure from a business. Whether and how the Fourth Amendment applies to those routes was not before the Court. An agency considering them should obtain its counsel’s view, and a business asked to disclose voluntarily should consider its own obligations to customers and patients before agreeing.

Training and audit

A policy reaches investigators through training and is tested through audit. Training needs to explain the categories as well as the rule for cell-site records, because new data sources keep arriving. Audit can sample requests and check that the process sought matched the category of record, that exigency claims were documented and that warrants were obtained where the policy required them.

The audit also protects cases. Evidence obtained on deficient process can be challenged later. An agency that finds a pattern of deficient requests in its own audit can correct it before a court does.

Open questions as policy decisions

The Court left open real-time data, tower dumps and other location-revealing records. For agencies, an open question can be handled by litigation or by policy. An agency can decide to seek warrants for categories the Court did not reach, to reduce the risk of suppression later. That choice trades investigative speed for legal durability.

The decision’s narrowness made that trade a recurring one. Each new data source, from vehicle telematics to app location histories, could raise the question again. Agencies that set a policy and documented its reasoning could answer consistently. Those that decided case by case risked inconsistent results across investigators.

Objections, and where the argument is weakest

Four objections test the argument that the decision made legal process a classification problem.

The objection that the decision was narrow

The Court called its decision narrow. On that view, holders other than carriers had nothing to classify, because the Court expressly left their records alone.

The Court left those records alone for the time being, and it said so in the language of open questions rather than settled answers. Its reasoning turned on features of data, not on the identity of the holder. A holder that waited for a later case to tell it which of its data sets fell on which side would be classifying under pressure. As the hindsight below shows, a later case did extend the reasoning.

The objection that this is the government’s problem

A second objection is that the Fourth Amendment binds the government, not private holders. A bank or hospital that complies with facially valid process has done its duty, and suppression is the government’s risk.

As to legal liability, that may be largely right, and it is a question for counsel in each case. It does not answer the operational and reputational questions. A holder that discloses a customer’s or patient’s location history on deficient process may face questions from customers, regulators or courts about its own handling. And a holder that cannot classify its data cannot decide when to push back, seek clarification or give notice where permitted.

The objection that holders will over-refuse

Another objection is practical. If holders classify aggressively, they may refuse or delay demands that the law permits, slowing investigations. Emergencies could be harmed by a legal process team waiting for counsel.

The decision itself answers part of this. It preserved exigency, and a well-run legal process function has an emergency path with defined criteria and fast escalation. The rest is answered by the Court’s default: subpoenas suffice in the overwhelming majority of cases. A classification guide that reflects that default narrows escalation to the records that need it. Consistency, not caution, is the aim.

The objection that the dissents were right about uncertainty

Justice Kennedy’s dissent warned that courts and officers would have to guess how much information could be requested before a warrant was required. That uncertainty is real, and it cuts against expecting holders to classify with confidence.

The objection supports the argument more than it undermines it. Uncertainty is the reason a holder needs an inventory and a policy. Where the law is unsettled, consistent internal classification, documented with counsel’s reasoning, is the best a holder can do. It is also what a court or regulator would later look for.

Justice Kennedy’s dissent made a further point that bears on requesters. He wrote that, by invalidating the government’s use of court-approved compulsory process, the Court called into question the subpoena practices of grand juries, legislatures and other investigative bodies. Whether that fear was borne out depends on later cases. For an agency that issues subpoenas, the warning is a reason to review which of its demands seek records with the features the Court described. It can then decide in advance what process it will use for them.

Checks for a records custodian and an investigating agency

The four parts translate into checks that apply on both sides of a demand.

  • Inventory: list every data set your organization or its vendors hold that records location, and note for each how detailed, continuous and long-retained it is.
  • Retention: for each location data set, confirm a business reason for its retention period, and shorten periods that have none.
  • Intake: give the legal process team a classification guide, reviewed by counsel, that maps each data set to the process the organization requires before disclosure.
  • Vendors: require vendors holding your location data to notify you of demands where the law permits, and to follow your classification guide.
  • Health and finance overlays: for health data, check each disclosure against 45 CFR 164.512(f); for federal demands for financial records, against the Right to Financial Privacy Act.
  • Agency policy: for agencies that request records, set a written policy on what process you will seek for each category the courts have left open, and record exigency facts when you rely on them.

The checks do not require predicting the next case. They require an organization to know what location data it holds and to decide in advance how it will answer a demand for each kind. The 2018 decision made that knowledge necessary. The 2026 geofence decision made it urgent.

The checks read differently by sector. A bank should start with its digital channels and fraud tools, where device and location data accumulate. A health system should start with patient apps, remote monitoring and access logs. A state or local agency should start with transit, toll, vehicle and app records, and with the policy that governs requests from police. A carrier or app provider is already on notice; its task is to keep the classification current as its products change.

With hindsight

Questions the Court left open in 2018 returned through geofence warrants. A geofence warrant asks a technology company for data on devices present in an area during a period, rather than records on a known person. In a robbery case from Virginia, police obtained such a warrant to Google in June 2019 for location data near a credit union that had been robbed. A divided Fourth Circuit panel held there was no search, and the full court, sitting en banc, affirmed in a one-sentence order after dividing evenly on the search question.

The Supreme Court granted review on January 16, 2026, limited to whether the police violated the Fourth Amendment in obtaining the location data. It heard argument on April 27 and decided the case, No. 25-112, on June 29, 2026. Justice Kagan wrote for the Court, joined by Chief Justice Roberts and Justices Sotomayor, Kavanaugh and Jackson. The Court held that police conducted a Fourth Amendment search when they acquired the location data from Google, because an individual has a reasonable expectation of privacy in cell-phone location information.

The Court relied directly on its 2018 reasoning. It said everything that decision relied on applied as well or better to Google’s Location History data. That data, the Court said, could pinpoint location within about twenty meters and recorded it every two minutes or so. The Court rejected the argument that obtaining only a short period of data avoided a search, and rejected the argument that users had voluntarily shared the data by turning the feature on. It said an app-by-app, feature-by-feature method of granting Fourth Amendment protection misapprehends the nature of modern cell-phone use.

The Court did not decide whether the warrant was valid. It left to the Fourth Circuit whether each step of the multi-step warrant satisfied the requirements of particularity and probable cause, and vacated and remanded. Justice Jackson concurred and Justice Gorsuch concurred in the judgment. Justices Alito and Barrett filed dissents; Justice Thomas joined part of Justice Alito’s, and Justice Barrett joined parts of it. The Court’s judgment issued on July 31, 2026, sending the case back to the Fourth Circuit with the warrant’s validity still to be decided.

The opinion recorded how far the practice had grown. It said Google received its first geofence warrant in 2016, received 982 two years later and more than 11,000 two years after that. It also noted that in July 2025 Google began storing Location History on users’ devices and represented that it could no longer respond to geofence warrants seeking that data.

What held up from June 2018 was the reasoning about features of data: detail, continuity, retrospection and involuntary collection. The 2026 Court applied it to a different holder and a different kind of request. In our reading, the narrowness the 2018 Court claimed did not last. We would expect holders that had classified their location data by those features to have been better placed for the extension. Google’s own change pointed to the other answer: data a holder does not keep cannot be demanded from it. For every other holder, the 2026 decision added another category to classify.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • Fourth Amendment
  • location data
  • Stored Communications Act
  • legal process
  • geofence warrants