A “helpful” restaurant robot turned into a flailing, dish-launching hazard in California—raising an uncomfortable question about how quickly businesses are replacing human judgment with automated systems.
Quick Take
- A humanoid service robot at a Haidilao hot pot restaurant in Cupertino (near San Jose) struck a table and sent tableware flying during a routine performance.
- Three employees physically restrained the robot as it continued erratic movements; no injuries were reported.
- Haidilao’s explanation centers on positioning error—placed too close to a table at a customer’s request—rather than a technical malfunction.
- The viral video amplified public skepticism about customer-facing automation in tight, unpredictable real-world settings.
What happened inside the Haidilao dining room
Restaurant patrons in Northern California watched a planned robot “performance” go sideways when a humanoid unit began striking a nearby table and scattering chopsticks, condiments, and dishes across the dining area. Reports place the incident at a Haidilao hot pot location in Cupertino, often described as near San Jose, and the video circulated widely in mid-March 2026. Staff members rushed in as the robot kept flailing, turning a novelty feature into a safety concern.
The clearest consistent details across coverage are straightforward: the robot started a programmed routine, made contact with a table, and continued moving erratically as items flew. One employee initially tried to intervene but couldn’t stop it, and two additional workers joined to physically restrain the machine. No injuries were reported, and the situation ended only after human staff overpowered the device and regained control of the immediate area.
Company response: “Not a malfunction,” but a setup problem
Haidilao representatives disputed the popular “berserk robot” framing by pointing to a positioning mistake. According to the company’s account relayed in U.S. media coverage, the robot was placed too close to a table at a customer’s request, leaving insufficient space to operate safely during the routine. That explanation matters because it shifts the blame from defective technology to operational decisions—where management and staff training, not marketing hype, determine whether automation stays contained.
The public-facing dispute—glitch versus human setup error—also reveals the limits of what outside observers can verify from a short viral clip. The manufacturer was not identified in the available reporting, and the company did not provide a technical breakdown of what sensors, collision-avoidance, or emergency-stop systems were in use. Without those specifics, audiences are left with what they can see: a machine in a crowded dining room that required a hands-on tackle to stop.
Why this viral moment lands differently in 2026
Americans have spent years being told that automation is inevitable, efficient, and “safer” than people. But viral scenes like this cut through corporate messaging because they show what happens when real life doesn’t match the lab conditions. Tight restaurant layouts, unpredictable customer behavior, and employee pressure to keep service moving can expose weak points fast. Even with no injuries, the incident underscores that “convenience tech” can become a physical risk in seconds.
The practical lesson: accountability doesn’t disappear when humans are replaced
Automation doesn’t eliminate responsibility; it redistributes it. If the company’s account is accurate, the problem started with a decision to position the robot too close to a table to satisfy a customer request. That points to the need for firm safety rules that employees can enforce without being second-guessed by “the customer is always right” culture. If the robot truly malfunctioned, then the accountability shifts toward engineering safeguards and clear shutdown procedures.
Hot Pot, Hot Mess: Service Robot Goes Berserk In San Jose Dining Room, Must Be Tackled By Staff https://t.co/WsIx1y6T9L
— zerohedge (@zerohedge) March 25, 2026
Either way, the event is a reminder that when businesses deploy machines in public spaces, the standard can’t be “it usually works.” Customers don’t sign up to be beta-testers in a dining room, and workers shouldn’t be forced into unsafe improvisation to protect the public and the brand. With limited reporting beyond March 2026 coverage, the best takeaway is simple: human oversight—and the authority to stop the show—still matters.
Sources:
AI Robot Incident Sends Tableware Flying at Restaurant
Out-of-Control Robot Destroys Dishes as Staff Try to Stop It
Humanoid Robot Goes Haywire at Hot Pot Restaurant in California, Sending Tableware Flying
