TL;DR: Proof News and 404 Media tracked 21 security robot deployments in the US since 2015. At least 13 of those programs have ended. Knightscope, the Silicon Valley firm with the most contracts, has also suffered the most cancellations. New York City's Times Square pilot expired in 2024. Dublin, Ohio pulled its "DubBot" robot after less than 10 months of a planned two-year pilot. San Antonio's airport declined a $21,000 one-year lease after a one-month test. Knightscope's CEO just bought Event Risk LLC and is pairing robots with human guards. The pivot is the story [1][2].
The Roboguard Pilots Are Not Sticking
Robotics companies have sold the pitch that video-camera-toting security robots can deter and detect crime. The market is reconsidering. Proof News found evidence of at least 21 security robot deployments since 2015 and confirmed that at least 13 of those programs have ended. Knightscope, based in Silicon Valley, secured the most contracts and absorbed the most cancellations [1][2].
New York City installed a Knightscope robot on the overnight shift at the Times Square subway station during Eric Adams' tenure. The pilot expired in 2024. City leaders did not respond to Proof News' questions about why the robot was not renewed. By the end of its assignment it was reportedly gathering dust in an empty storefront [1][2].
Dublin, Ohio, pulled the plug on its "DubBot" Knightscope robot in May after less than 10 months of a planned two-year pilot. City spokeswoman Robyn Gray told Proof News the robot "did not fully meet our operational needs." The pilot identified no criminal incidents, no tickets, and no arrests. Salem, Oregon ended a three-month Daxbots pilot in April 2026. Spokeswoman Nicole Miller cited "no additional funding." A Boston Dynamics "Digidog" lease at the NYPD was ended in 2021 after public backlash. Representative Alexandria Ocasio-Cortez called those robots "surveillance ground drones" "deployed for testing on low-income communities of color" [1][2].
Why the Pilots Keep Failing
The consistent finding is that the robots are not built for the work. San Antonio International Airport spokesperson Ana Flores told Proof News the airport's 2024 Knightscope robot struggled with badge scanning, communication, and navigation. Airport leadership declined a $21,000 one-year lease after a one-month test [1][2].
Sheila Sparks, owner of Sparks Protection Services near Columbus, framed the gap directly. A human guard handles the work a robot cannot: "A robot can't do that." Missy Cummings, a robotics professor at George Mason University, put the broader problem in one sentence. "AI and robots for security have had a long, problematic history" [1][2].
Andrew Ferguson, a law professor at George Washington University, was more direct. He called security robots "an alluring form of surveillance as a service" and described the machines as a symbol rather than a tool. The operational summary was blunt: "Robots are a form of security theater, more optics than results." Steve Rosta, co-owner of Direct Approach Security Services in Ohio and a retired highway patrol officer, drew the line the same way: "Boots on the ground is always better" [1][2].
Knightscope Buys a Human-Guards Firm
Knightscope CEO William Santana Li announced earlier in 2026 that the company had purchased Event Risk LLC and would combine robots and AI with human guards. The bet is the $230 billion annual physical security market. Knightscope has run an accumulated deficit since its 2013 founding and carries $273 million in debt per its most recent SEC quarterly filing [1][2].
Li framed the move as complementarity. His pitch: a guard-plus-robot team can do what neither can alone. The private security labor market supports the read. US private security guards earn about $40,000 a year per the Bureau of Labor Statistics, roughly half a police officer's salary. There is room to build a hybrid service that is cheaper than sworn officers and more useful than a robot [1][2].
A Decade of Mishaps
The Knightscope track record is the context the new pitch lands against. In 2016, a Knightscope robot at a Silicon Valley shopping center rolled over a 16-month-old boy's foot. The company called it a "freakish accident" and said the robot "veered" to avoid the child, who ran backwards into it. In 2017, a Knightscope robot fell into a Washington, D.C. fountain. The company tweeted: "BREAKING NEWS: 'I heard humans can take a dip in the water in this heat, but robots cannot. I am sorry.'"[1][2]
The broader market has the same shape. Daxbots deployed to IDEA Tempe, a mixed-use campus in Tempe, Arizona. The contract ended. Adam Pioth, who posted a video of one Daxbot in operation, told Proof News the bots would "stop, look at me, and then just keep going really slow" when a person crossed their path. Daxbot pivoted. The company now sells to at least nine cities for ADA sidewalk compliance assessments, a different business than the one its original marketing promised [1][2].
The Federal Import Ban Reshapes the Market
A new federal rule against foreign-made mobile robot imports may be the structural reason Knightscope is back in the deal-making mood. The Federal Communications Commission banned the imports on national-security grounds. US-only deployments open space for domestic vendors. Boston Dynamics, majority-owned by Hyundai since 2021, has stayed in the news. The Atlas humanoid delivered the match ball at a game in July 2026. Two World Cup stadiums recently deployed Boston Dynamics robots [1][2].
The FCC action does not change the pilots. What it does is keep foreign competitors out while domestic ones prove out their model. Knightscope's pitch now is that the model is a robot plus a human, not a robot alone. The pivot is the tell. When a robotics firm buys a guard company, the robot is the side product. For a privacy audience the lesson is the same one the broader surveillance-tech wave keeps teaching: the device is the surface. The data it collects is the product. Read the police drone surveillance vessel for the pattern at a larger scale.
What to Watch
Knightscope's first quarterly after the Event Risk integration. The accumulated deficit and $273 million debt load are public. The new line item will tell whether the hybrid model can sell [1][2].
Salem, Oregon's next move. The Daxbots pilot ended because there was "no additional funding." Whether the city returns to the same vendor or a different one is the test of whether the cancellation was about this robot, or this category [1][2].
The Boston Dynamics Atlas rollout at the World Cup stadiums. Two stadiums have taken deliveries. The post-event reviews will say whether the robots worked or were the "optics" Ferguson described [1][2].
The next public-safety agency that buys the hybrid pitch. The pivot is a marketing claim. A municipal contract with a published RFP and a deployment report would let readers see what the "robot plus human guard" actually does [1][2].
Sources
- 404 Media, Rebecca Plevin and Varsha Bansal: The Roboguard Revolution Is Short-Circuiting (August 10, 2026). https://www.404media.co/the-roboguard-revolution-is-short-circuiting/
- Proof News, Rebecca Plevin and Varsha Bansal: The Roboguard Revolution Is Short-Circuiting (August 10, 2026). https://www.proofnews.org/the-roboguard-revolution-is-short-circuiting/