Switzerland’s Leibstadt Nuclear Power Plant (KKL) has adopted quadruped robots from DEEP Robotics for internal inspections. The announcement, made in July 2026, marks one of the first operational deployments of this kind in Europe and opens a concrete path toward digital maintenance in nuclear facilities.
KKL is Switzerland’s highest electricity-producing nuclear plant, capable of meeting roughly one-seventh of the country’s national demand. A facility of this scale requires continuous, high-frequency monitoring. DEEP Robotics’ robots follow pre-planned routes alongside technicians, take instrument readings, capture thermal images, and detect acoustic anomalies or gas concentrations — all autonomously, without interruption, 24 hours a day. The collected data is synchronized to a local management system hosted within the plant’s secure network, in compliance with Swiss data sovereignty and cybersecurity regulations.
The most immediate benefit concerns personnel safety. In certain sections of the plant, staff are required to wear heavy protective gear, and repetitive checks in confined environments mean prolonged exposure. The robots take on these tasks, freeing human operators to focus on more complex assessments. Some areas are structurally compact — ducts, crawl spaces, gaps between machinery — and difficult to access with conventional equipment. Quadruped locomotion, with the ability to climb slopes of up to 45 degrees and navigate tight spaces, enables continuous monitoring even where a wheeled cart or drone would struggle to operate.
The industrial context is clear: mobile robotics applied to nuclear settings is no longer a laboratory experiment. Boston Dynamics has already supplied its Spot robot to operators such as Ontario Power Generation, Talen Energy, and Duke Energy’s Oconee Nuclear Station for remote inspections and decommissioning support. What sets the KKL case apart is its integration with an in-house digital maintenance management system — an approach designed to close the loop between anomaly detection and technical response, without routing data through external platforms.
DEEP Robotics is a Chinese startup, and its presence at a European nuclear plant is no coincidence. It fits squarely within Beijing’s strategy of promoting its so-called “next-cycle” industries abroad: artificial intelligence, robotics, and advanced pharmaceuticals. The Leibstadt plant is, in this sense, also a showcase. But beyond the industrial narrative, the technical case remains solid: a robot conducting high-frequency checks in difficult-to-reach areas, without stopping at night or on weekends, reduces the variability inherent in manual inspections and increases the density of data available to maintenance managers.
The trajectory emerging from Leibstadt suggests that over the coming years, mobile robots will become standard components of nuclear maintenance — not experimental tools. The real challenge will not be technical, as quadruped locomotion is already mature, but organizational: integrating robot-collected data into plant decision-making workflows, with clear protocols defining how and when an anomaly flagged by a machine triggers a human response.
