Walk through almost any modern car plant or aerospace hangar and you will find two kinds of robots. The first kind builds things: it welds, rivets, paints and lifts. The second kind, far rarer, checks whether the first kind did its job correctly. That second job has always been harder to automate, because a robot arm that is good enough to move a welding torch is rarely precise enough to measure a part to a fraction of a millimeter. On November 5, 2026, ZEISS announced a machine designed to close that gap. According to ZEISS’s press release, the new ZEISS TALOS metrology robot combines “metrology-grade accuracy” with the reach and flexibility of an industrial robot, and it does so without the external tracking cameras or stick-on reference markers that similar systems usually depend on.
What ZEISS actually launched
TALOS is not a single arm but a small family of components that work together. ZEISS describes four pieces: the ZEISS TALOS MR metrology robot, the compact ZEISS TALOS AX blue-light sensor that rides on it, a metrology-grade ZEISS TALOS RT rotation table that turns the part being measured, and a safety concept called ZEISS TALOS Safeguard. Together they form what the company calls an integrated measurement unit.
The headline number is accuracy over distance. ZEISS says the system delivers measurement uncertainty below 100 micrometers (less than a tenth of a millimeter) across a reach of 3.30 meters. That reach matters because the parts it targets are big: car bodies, truck cabins, aircraft structures and jet engines. Holding sub-0.1 mm uncertainty over more than three meters, with the robot itself positioning the sensor “precisely and repeatably,” is the core technical claim of the launch.
Dr. Jan Thesing, Head of Business Unit 3D Optical Systems at ZEISS, framed it as a category shift, saying in the release that “metrological capabilities are integrated directly into the system, while the speed, range and flexibility of robotic kinematics are retained.”
Why “no markers, no tracking” is a big deal
Ordinary industrial robots are repeatable, meaning they can return to the same spot again and again, but they are not necessarily accurate in absolute terms. Joints flex, arms warm up, and small errors add up along the length of the arm. To use a standard robot for precise optical inspection, manufacturers typically bolt on an external tracking system, such as cameras that watch the sensor, or they cover parts and fixtures with reference markers so software can stitch scans together.
Both workarounds add cost, setup time and points of failure. Tracking cameras need a clear line of sight to the sensor, which is a problem when the sensor has to reach inside a wheel arch or around the back of an engine casing. ZEISS says TALOS removes that dependency, which in turn lets it measure “areas that are difficult to access with conventional line-of-sight measurement systems.” For a factory, that means fewer accessories to calibrate and maintain, and fewer blind spots on the part.
One sensor, two kinds of measurement
Optical inspection has long forced a trade-off. Fast full-surface scanning is great for building a digital twin and seeing overall shape deviations, but it is less suited to pinpoint feature checks such as the exact position of a bolt hole or the edge of a panel. Those features have often required a separate, slower process.
The TALOS AX sensor is ZEISS’s answer. According to the company, it uses two 24-megapixel cameras capturing at 100 Hz, or up to 200 Hz in a fast-scan mode, with a 500 mm wide measurement volume. Four multidirectional “feature flashes” light the part from different angles so the same sensor can capture both broad surfaces and fine features, including holes, edges, bolts and threads, in a single automated workflow.
Built for the factory floor, not just the lab
Precision instruments traditionally live in climate-controlled measuring rooms. ZEISS says TALOS is built to withstand common production conditions, naming temperature variation, vibration and dust, so it can work “near production” as well as in dedicated metrology rooms.
Software is the other half of the pitch. Measurement jobs can be prepared offline from a part’s CAD model, and the company’s ZEISS INSPECT software automatically calculates the measurement positions and the robot paths needed to reach them. ZEISS highlights one-click programming and streamlined setup, and says users can run full analyses, generate reports and feed results into existing quality processes. The practical promise is that adding a new part to the inspection schedule should take hours of planning rather than days of manual robot teaching.
Who it is for
ZEISS lists automotive, aerospace and commercial vehicle manufacturing as the target industries. Example jobs include inspecting car bodies and hang-on parts such as doors and hoods, large single-piece “giga-casting” components, aerostructures, jet engines, truck cabins and weld seams. The company also notes that virtual tools can evaluate functional dimensions such as gap size and flushness, the small details that decide whether a car door lines up neatly with the body panel next to it.
What it means
- Inspection can keep pace with production. As carmakers move to bigger cast parts and faster model changes, measuring more parts, more often, without slowing the line is a competitive advantage. A self-positioning metrology robot is built for exactly that.
- Less infrastructure, fewer failure points. Removing external trackers and markers simplifies cell design and cuts the setup work that often makes automated inspection expensive to deploy.
- Richer data for digital twins. Capturing full surfaces and precise features in one pass gives engineers a more complete picture of each part, which supports root-cause analysis when something drifts out of tolerance.
- A signal for the wider robotics market. TALOS is part of a broader trend in which robots are judged not just on strength and speed but on how well they sense and measure the world around them.
ZEISS’s release does not give pricing or a specific shipping date, so buyers will need to talk to the company directly about availability, and real-world performance claims will be tested as the first systems go into service.
The bottom line
Quality control is one of the least glamorous parts of manufacturing, and one of the most expensive to get wrong. With TALOS, ZEISS is betting that a robot can be both the flexible mover and the trusted measuring instrument, without a scaffolding of cameras and markers around it. If the sub-100-micrometer figure holds up across a 3.3-meter reach in real factories, the robots that check the work may finally become as common as the robots that do it.