They Know Where You Parked Last Tuesday: How 250 Million License Plate Scans Per Month Turned Your Commute Into a Police Database

A mother was driving her minivan with her young children in the back seat when police officers swarmed the vehicle, drew their weapons, and ordered the entire family out of the car and face-down onto hot asphalt at gunpoint. The officers believed they had located a stolen motorcycle. They had not. An automated license plate reader had misread the family's plate number, and the system flagged their minivan as a match. The city of Aurora eventually paid the family $1.9 million to settle the lawsuit. The camera that triggered the stop was never audited. The algorithm that generated the false match was never reviewed. And the company that built the system continued expanding into thousands of additional American cities.

That company is Flock Safety. And the odds are high that its cameras are already recording your vehicle, right now, in your town, as you read this.

What the Cameras Actually Capture

Flock Safety's core product is the Falcon, a solar-powered camera mounted roughly seven feet off the ground on public roadways, neighborhood entrances, and commercial parking lots. The Falcon uses infrared flash technology to photograph vehicles at speeds up to 100 miles per hour in any lighting condition. But describing Flock as a license plate reader fundamentally misrepresents what the system does.

Each camera captures far more than an alphanumeric plate number. Onboard artificial intelligence generates what the company calls a "Vehicle Fingerprint," a detailed profile that includes the vehicle's make, model, color, body type, and distinguishing features such as roof racks, bumper stickers, and body damage. The system can identify and track a vehicle even if the license plate is removed, intentionally covered, or misread by the optical character recognition software.

Every scan is timestamped, geolocated, and uploaded to a centralized cloud database. That database is searchable by any authorized user across any participating jurisdiction in the country. As of 2026, the network operates in more than 6,000 cities, processes over 250 million plate scans per month, and retains the data for periods ranging from 30 days to indefinitely, depending on the local contract.

Your car was photographed this morning. It was photographed yesterday. The system knows what time you left your house, which route you took to work, how long you stayed, and which direction you drove when you left. It has known this for months.

The Business Model: You Pay, They Own

Traditional surveillance equipment is purchased by police departments as a capital expense. The city owns the hardware. The city's IT department can inspect the software. The city controls the data. Flock Safety operates on a fundamentally different model.

Flock retains ownership of every camera. Municipalities, homeowners associations, and retail chains like Home Depot and Lowe's pay a recurring annual subscription fee, typically between $2,000 and $3,500 per camera per year. This Hardware-as-a-Service model guarantees the company continuous revenue extraction from taxpayers while legally shielding the equipment from local ownership or independent technical review. The user agreements explicitly prohibit municipalities from decompiling, reverse engineering, or tampering with the devices.

The city pays for the cameras. The city cannot open them. The city cannot audit the code that decides whether your plate is a match. The city cannot verify whether the data is being shared with agencies it never authorized. The company retains total leverage over public safety infrastructure that taxpayers fund but do not control.

This model has made Flock Safety extraordinarily valuable to investors. In early 2025, the company closed a $275 million funding round that pushed its valuation to $7.5 billion. A subsequent $208 million raise in 2026 elevated it to $8.4 billion. The primary architect of this capitalization is Andreessen Horowitz, one of the world's most influential venture capital firms, which has designated Flock as a cornerstone of its "American Dynamism" portfolio. Additional investors include Tiger Global, Matrix Partners, Founders Fund, and Meritech Capital. Annual recurring revenue reportedly exceeds $300 million.

The venture capital mandate for exponential growth dictates a hyper-aggressive go-to-market strategy. Between the company and its primary investor, nearly $92 million was expended in political lobbying across the United States in a single year to secure municipal contracts and influence state privacy legislation. In Cleveland, Ohio, a city council member resigned his position only to take a public affairs job at Flock Safety days later. Shortly thereafter, an emergency ordinance was introduced to award the company a sole-source contract, bypassing competitive bidding entirely. In Orlando, Florida, nearly 100 cameras were installed on public roadways without notifying the county government.

The Efficacy Myth: 10% of All Crime, or a Marketing Slide?

Flock Safety's sales literature frequently cites an extraordinary claim: that its technology assists in solving more than 10% of all reported crime in the United States and can produce 100% homicide clearance rates in specific test cities like Tulsa, Oklahoma. These figures are generated internally by the company. They have never been published in a peer-reviewed journal. They have never been subjected to independent methodological review.

The independent research tells a different story. Dr. Cynthia Lum and Dr. Christopher Koper of George Mason University's Center for Evidence-Based Crime Policy have conducted extensive randomized controlled trials and longitudinal studies spanning multiple jurisdictions. Their consistent finding is that automated license plate readers provide little clear evidence for general crime deterrence in patrol settings. While ALPRs do generate an increase in the raw number of "hits" regarding stolen vehicles, this data influx does not proportionally translate to higher arrest rates, improved case clearances for broader crime categories, or a measurable reduction in violent crime.

A detailed case study analyzing ALPR implementation in the Tigerland area of Baton Rouge, Louisiana, examined crime rates fifteen months before and fifteen months after installation. The study found negligible changes in crime type distribution and concluded that the cameras had no statistically significant effect on violent crimes such as assault, battery, homicide, or robbery.

The Berkeley Police Accountability Board in California went further. Their review found that in several communities, the introduction of ALPRs was actually correlated with increases in vehicle theft and lower overall crime clearance rates. The board flagged a specific instance where Flock Safety claimed its services were responsible for a 33% decrease in motor vehicle thefts in Vallejo, California, but upon examination, the data the company cited came from years before its cameras were ever installed.

The False Positive Problem

The Aurora gunpoint stop was not an anomaly. Optical character recognition systems are inherently prone to errors. They misread plate numbers. They confuse states with similar license plate designs. They flag benign vehicles as stolen. Every false positive generates an alert that an officer treats as a confirmed match, and the consequences for the innocent driver are immediate and potentially lethal.

Similar incidents of ALPR-triggered false alerts leading to wrongful detentions at gunpoint have been documented across California, the Carolinas, and New Mexico. In each case, the system operated exactly as designed. The camera captured an image. The algorithm generated a match. An officer responded with force. The person in the vehicle had done nothing wrong.

The racial dimensions of these errors compound the danger. Data audited from Oak Park, Illinois, revealed that 84% of drivers stopped in Flock-related traffic incidents were Black, despite Black residents comprising only 19% of the local population. The Electronic Frontier Foundation reviewed license plate surveillance logs from various agencies and found that the racially coded language entered into Flock's search justification fields mirrored long-standing patterns of discriminatory policing, utilizing terms like "furtive movements" or "suspicious behavior" to target historically marginalized communities.

Oak Park's village board examined the data and reached a conclusion the company's marketing materials never acknowledge: the cameras had not meaningfully solved a single local crime since 2022. In August 2025, the village voted to terminate its Flock Safety contract.

What the Cameras Are Becoming

The Falcon plate reader was the beginning. It is no longer the product.

Flock's current hardware ecosystem includes the Condor, a pan-tilt-zoom video camera that operates on a cloud-based subscription and features "Guardian Mode," an AI system that automatically detects, zooms in on, and tracks both vehicles and pedestrians without human intervention. It includes the Raven, an acoustic sensor that detects gunfire and cross-references audio events with visual ALPR data to identify fleeing vehicles. And following its acquisition of the drone company Aerodome, it includes automated aerial drones that provide what the company markets as "Drone as First Responder" capability.

But the most consequential escalation is not hardware. It is a software platform called Flock Nova.

Nova consolidates license plate data, Condor video feeds, Raven audio alerts, third-party data sources, Computer-Aided Dispatch records, and police Record Management Systems into a single search interface. Marketed with the tagline "one click, one case solved," the platform allows investigators to map relationships between individuals, vehicles, and historical data patterns across a unified intelligence dashboard. In April 2025, investigative reporting by 404 Media revealed that the beta version of Nova was ingesting data obtained from large-scale hacks and security breaches on the dark web to augment its intelligence profiles. Flock announced it would stop using the illegally obtained data. The architecture that enabled the ingestion remains intact.

In 2025, Flock added "FreeForm," a natural language processing layer that eliminates the need for structured database queries. Officers no longer need a plate number. They type conversational descriptions into the system: "white SUV with a racing stripe." "Red pickup truck with a dog in the back." The AI searches billions of stored images and returns matches. The technical barrier to mass surveillance has been reduced to a sentence typed into a search box.

The system also includes a "Convoy" detection algorithm. An investigator inputs a single vehicle, and the algorithm sifts through the network's stored data to identify other vehicles that frequently appear near the target across multiple cameras within a specific time window. The system ranks companion vehicles by recurrence and proximity, determining whether a car is consistently leading or following another. It automates the mapping of personal associations and travel networks without requiring a warrant or reasonable suspicion of a crime.

This is no longer a license plate reader. It is a privately owned intelligence fusion center with a $2,500-per-year subscription fee, funded by the same taxpayers it tracks.

Next in this series: The database that lets cops track ex-girlfriends, feds hunt immigrants in sanctuary cities, and corporate employees watch the preschool daycare camera for "sales demos." The people with access to your location history are the ones you should be most afraid of.

Disclaimer: This article is for informational and journalistic purposes only and does not constitute legal advice. References to corporate valuations, investment rounds, municipal contracts, and academic research are provided for public transparency and reader reference. Guerrilla News is an independent journalism outlet and is not affiliated with any company, government agency, investor, or legal proceeding discussed in this report.

Sources

1. Flock Safety, Wikipedia
2. Flock Safety, "Flock Safety Raises $275M in New Funding at $7.5B Valuation"
3. "The Case Against Flock Safety," DeFlockOKC
4. "How a $7.5 Billion Surveillance Company Built America's Largest Warrantless Tracking Network," MyPrivacy Blog
5. Flock Safety, Products: Cameras, Trailers, LPR, Drones & Software
6. "Flock Safety Secures $275M in Funding at $7.5B Valuation," Built In
7. Flock Safety 2026 Company Profile, PitchBook
8. "Investing in Flock Safety," Andreessen Horowitz
9. "Who Owns Flock Safety Company?" Business Model Canvas Research
10. Lobbyist Reports Search, Georgia Government Transparency and Campaign Finance Commission
11. License Plate Readers (LPR) Cameras, Flock Safety
12. "What the Police Actually See When They Run a Flock Search," DeFlockILM
13. "Flock Safety Launches New AI-Powered Tools," Police1
14. "Flock Nova: Smarter Investigations, Faster Case Wins," Flock Safety
15. "Flock Safety Reveals the Most Expansive AI and Data Analysis Toolset for Law Enforcement," GlobeNewswire
16. Cynthia Lum, George Mason University, Center for Evidence-Based Crime Policy
17. "The Adaptation of License-plate Readers for Investigative Purposes," ResearchGate
18. CAPER Fact Sheet 18, Louisiana State University
19. "Measuring the Cost-Effectiveness of ALPRs in Policing," ResearchGate
20. "Berkeley Memo Lays Out New Concerns Over Flock Cameras," KALW
21. BCOS Resolution 2026-02 DRAFT with Memo, City of Bloomington, Indiana
22. "License Plate Surveillance Logs Reveal Racist Policing Against Romani People," Electronic Frontier Foundation
23. Automated License Plate Readers, Village of Oak Park