Cities that have dropped Flock Safety

Between mid-2025 and mid-2026, ending a Flock Safety contract went from a rare municipal decision to a weekly news item. Newsweek, mapping decisions compiled by the crowdsourced ALPR-tracking project DeFlock, reported in late July 2026 that communities in 23 states had rejected, canceled, or deactivated Flock license plate readers in 2026 alone, with California and Wisconsin leading at seven apiece.¹ DeFlock’s broader tally at that point counted 47 jurisdictions that had canceled contracts, 25 that had voted down proposed agreements, and 22 that had switched cameras off.¹

The overall totals depend on who is counting and what counts. The Washington Examiner put the number of communities that have walked away from the company since 2021 at roughly 98;² Futurism counted at least 54 cities since the start of 2026.³ Some tallies include cities that rejected a proposal before any camera went up, others only terminated contracts. The direction, however, is consistent across every count, and it coexists with continued growth elsewhere: Flock’s network still spans well over 100,000 cameras nationwide,¹ many of which appear on our state-by-state camera index.

The tracker below is deliberately narrower than the headline numbers. It is the subset we could verify: each row is one municipality, county, or agency whose decision we confirmed against a specific news report or public record, linked in the row. Where an aggregator listed a city but we could not locate direct sourcing, we left it out.

Last reviewed Aug 8, 2026 · 14 verified entries, newest first

Action distinguishes five outcomes: canceled (terminated an active contract), non-renewal (let a contract expire or declined to extend it), paused (suspended cameras pending review), rejected (voted down a proposed contract), and deactivated (turned cameras off without a formal termination). Dates are decision dates as reported, not the dates cameras physically came down.

This list is not exhaustive. It includes only decisions we have verified against a linked report or record; aggregate counts cited above run higher. Know of a decision we missed, or an error in a row? Send a correction.

Accuracy findings. The most-cited single document of 2026 is the Los Angeles Police Department Inspector General’s audit released July 11, which examined two months of stolen-vehicle alerts from the department’s Flock cameras and found 161 vehicles incorrectly flagged as stolen in August and September 2025 — a 32.3 percent false-positive rate. The inspector general warned that false alerts create “the risk of unnecessary enforcement actions, including vehicle stops and wrongful detentions.” LAPD let its three-year contract expire three days later, on July 14, 2026.⁴ The ACLU’s national campaign against ALPRs cites misread plates and mistaken stops as a recurring pattern beyond Los Angeles.¹¹

Federal access to local data. A second cluster of departures followed audits and lawsuits documenting federal queries of locally collected plate data. In August 2025, the Illinois Secretary of State’s audit found Flock had given U.S. Customs and Border Protection access to Illinois camera data through a pilot program — contrary to a 2023 state law barring use of plate data for immigration or abortion enforcement — and Flock’s CEO acknowledged the company “didn’t create distinct permissions and protocols” to ensure local compliance, pausing its federal pilots nationwide.⁶ In February 2026, a class action filed in San Francisco alleged that law enforcement agencies from outside California searched SFPD’s Flock database more than 1.6 million times between August 2024 and February 2025, in violation of California’s ALPR statute; the complaint separately alleges, without quantifying, that officers ran searches on behalf of federal agencies.⁵ And in Dayton, Ohio, the police department found more than 7,000 immigration-related searches by outside entities against city policy; workers covered all 72 cameras with trash bags while the program was suspended.⁷

Transparency and vendor conduct. Some terminations cite the company’s own operations rather than surveillance policy. Cambridge, Massachusetts terminated its contract in December 2025 after discovering two cameras installed without the city’s knowledge following an earlier deactivation — a “material breach of our trust and the agreement,” in the city’s words — and Eugene, Oregon opened an inquiry after at least one camera remained active weeks after officials ordered all 57 turned off.⁸ Other councils, as the tracker’s reason column shows, cited cameras activated before approval or contract terms that changed after signature.

Cost — and second thoughts in both directions. Contract dollars figure in many debates, and not every departure sticks. Cleveland let its $250,000-per-year contract lapse on June 29, 2026, then reversed course two weeks later, approving a six-month, $125,000 extension with new conditions: no data sharing with third parties, a public transparency portal, quarterly reporting, and cut off access for the regional fusion center.¹⁰ Reversals like this are why our tracker records dated decisions rather than permanent statuses.

Flock’s stated policy is that plate data is deleted on a rolling 30-day cycle by default, with the actual window set by each contract and applicable state law, and that deletion is enforced by an automated cloud lifecycle policy the company describes as irreversible.⁹ On that policy, a canceled customer’s footage would age out within weeks of the last capture. Flock’s public documentation does not, however, spell out a distinct end-of-contract data disposition process beyond that rolling window.⁹

Reporting has documented gaps between cancellation decisions and cameras actually going dark. In Eugene, at least one camera was still logging plates weeks after the city ordered its network shut off, capturing roughly 8,500 plates in a 30-day span; in Cambridge, new cameras appeared after the city had deactivated its account.⁸ In Cleveland, the cameras kept recording after the contract expired, pending the council’s renewal vote.¹⁰ For readers of this tracker, the practical takeaway is narrow and factual: a decision date marks when a government chose to end a program, not necessarily when collection stopped.

Inclusion criteria. A row requires a specific municipality, county, or agency; a decision documented in a public record or a credible news report we link directly; and a decision date. We use the date the decision was made (a council vote, a termination notice, a contract expiration allowed to pass), not the date of implementation. Aggregator lists — DeFlock, Newsweek’s map, ACLU campaign pages — are used to find candidates, never as the sole source for a row.¹

Maintenance. We re-verify the table roughly monthly; the review date shown above the table changes only when that check happens. Corrections and additions are welcome via the contact details on our About page.

For context on what these cameras record and how the network operates, see What is a Flock camera? For the statutes many of these decisions turned on — retention caps, sharing restrictions, and audit requirements — see license plate reader laws, state by state. To see where cameras are documented today, browse cameras by state and city.