Within a single week in May 2026, the city councils of Cambridge, Massachusetts and Bandera, Texas voted to terminate contracts with two of the most widely deployed AI-powered public safety platforms in the United States. One city has a population approaching 120,000 and a globally recognized research university. The other has fewer than 900 residents. What they share is a growing refusal to renew contracts on a vendor’s terms, a pattern now playing out across dozens of American cities and raising a structural question the smart surveillance industry can no longer defer: is the problem the technology, or the relationship?
The Terminations Keep Coming
Cambridge voted 5-2-2 on May 19 to decommission SoundThinking’s ShotSpotter gunshot detection network, ordering all acoustic sensors removed within 90 days. The city had used the system since 2014. Bandera, a Central Texas town, voted 3-2 on May 12 to end a planned rollout of eight Flock Safety license plate reader cameras funded by a state grant, before a single camera went fully live.
Neither vote was isolated. Denver ended its longstanding relationship with Flock Safety when the city’s contract expired on March 31 after years of escalating tension between the mayor’s office and a city council that had vocally opposed renewal. The city replaced Flock with Axon, awarding a one-year, $150,000 contract, though security researchers noted that Axon’s system collects similar categories of data and the switch may not resolve the underlying concerns. Chicago did not renew its ShotSpotter contract in 2024, a deal previously valued at approximately $49 million. Austin canceled its Flock contract in 2025 after more than 30 community groups organized against it. Little Rock ended its ShotSpotter arrangement in February 2025. According to data tracked by the advocacy platform Stop Flock Safety, at least 53 jurisdictions across 20 states had deactivated or rejected Flock systems in the six months prior to May 2026.
The pace suggests something beyond case-by-case controversy. A structural shift in how municipalities are evaluating and terminating automated public safety technology contracts is underway.
Three Failure Modes, One Pattern
The cases differ in geography, population, and politics, but the terminations share a recognizable anatomy. Three failure modes recur.
Accuracy credibility gaps. The New York City Comptroller’s office found in a 2024 audit that between 80 and 92 percent of ShotSpotter alerts during the reviewed period did not correspond to confirmed shootings. In Cambridge, only 35 percent of ShotSpotter notifications over a decade were confirmed as actual gunfire, according to council documents. When cities pay for systems marketed on response speed and crime reduction, and the underlying detection rate looks like this in public, the value proposition becomes politically indefensible regardless of whether the company disputes the methodology.
Data governance failures. Flock Safety’s loss of contracts in jurisdictions with sanctuary policies accelerated dramatically after 404 Media revealed the company’s “National Lookup” tool had been used in more than 4,000 immigration-related searches between January and May 2025. A subsequent investigation found that Flock CEO Garrett Langley had stated publicly the company held no federal contracts, while public records obtained by 9NEWS revealed a Border Patrol pilot program operating through the Loveland, Colorado police department. Trust, once broken at the CEO level, rarely survives a procurement renewal.
Governance architecture mismatches. Denver’s situation illustrates a third failure mode that vendors tend to underestimate. The city council voted unanimously against renewing the Flock contract, yet the mayor’s office executed two temporary extensions unilaterally by keeping the contract value below the council approval threshold. The council eventually prevailed when the contract expired, but the episode burned through institutional goodwill and left the city actively soliciting competitors. The lesson is not that mayors support surveillance and councils do not. It is that deploying technology into contested governance terrain, without investing in the civic architecture to sustain it, creates contracts that are one political cycle away from termination.
What Vendors Are Getting Wrong
Flock Safety, which counts more than 6,000 law enforcement and corporate customers and was valued at approximately $7.5 billion in 2025, framed early criticism as coordinated attacks from activists seeking to defund police. In an email to the Staunton, Virginia police chief, CEO Garrett Langley wrote that “Flock, and the law enforcement agencies we partner with, are under coordinated attack.” Staunton Police Chief Jim Williams publicly disagreed, describing the community’s concerns as “democracy in action.”
That gap in framing is commercially consequential. Vendors who characterize municipal pushback as an external threat rather than a product signal are systematically misreading their own churn data. The cities canceling contracts are not ideologically uniform: Bandera, Texas is a conservative rural community. Cambridge is a progressive sanctuary city. Denver sits somewhere between. If the rejection vector cuts across political lines, the explanation is not activism. It is a trust deficit with a structural cause.
The core issue is that the surveillance technology industry sold its first generation of products on operational value, speed of deployment, and law enforcement partnership, and essentially outsourced public legitimacy to its police department clients. That worked while public awareness of data-sharing practices was low. It stopped working when the data flows became visible.
What a More Durable Model Looks Like
Denver’s replacement of Flock with Axon is instructive, even if imperfect. The new contract was structured to require city council approval, data access was restricted to agencies with a formal memorandum of understanding, and the mayor framed the switch explicitly around protections for immigrants and people seeking reproductive healthcare. Whether Axon sustains that political compact over time remains to be seen, but the posture reflects what the market is beginning to demand: vendors who treat data governance as a product feature, not a legal afterthought.
Flagstaff Mayor Becky Daggett, speaking to NPR after her city ended its Flock contract, articulated the threshold that vendors have consistently underestimated. “In the end, it was just clear that this wasn’t going to be a technology that was going to be well received or that we could continue to use,” Daggett said.
That framing does not describe a technology failure. It describes a consent failure. The technology may work as specified. The problem is that the city could no longer maintain the political and community conditions under which using it was acceptable.
The vendors most likely to retain and grow public sector contracts over the next procurement cycle are those investing in three capabilities that are currently underbuilt across the industry: transparent, auditable data access logs that municipalities can independently review; clear contractual prohibitions on secondary data use that are enforceable rather than merely stated; and accuracy reporting that is published rather than defended. As Kurrant has reported in its coverage of Greece’s national AI traffic enforcement network, deployments that publish camera locations, disclose accuracy metrics, and frame the program around behavioral change rather than revenue tend to hold public support at scale. The camera does not have to be invisible to work. It has to be trusted.
The cities ending these contracts are not saying they do not want automated public safety technology. Denver immediately replaced Flock with a different vendor. Cambridge’s police department cited 11 incidents where ShotSpotter detected gunfire that generated no 911 calls. The operational case for the technology is not what is being rejected. What is being rejected is the version of the technology that arrives with opaque data pipelines, contested accuracy claims, and a governance posture that treats public concern as a threat to be managed rather than a relationship to be maintained.
