Filing activity is led by a small group of automation and robotics companies, with the top-ranked assignee holding 235 records in this dataset compared with 128 at fifth place and 57 at tenth. That said, the top five assignees combined account for only 19.5% of the 4,624 records in scope, so no single company controls the field outright. A long tail of other filers, including specialist robotics and automation firms, makes up the remaining majority of activity.

The dominant class is B25J, covering manipulators and robots, which touches 51.4% of the 4,624 records in scope. G05B, control and regulating systems, is the next largest at 24.0%, followed by smaller but still active classes covering digital data processing, non-electric control, AI-based computing, additive manufacturing, image processing and plastics shaping. Because a single record can carry multiple IPC classes, these shares add up to well over 100% and should not be summed as if they were exclusive categories.

Filing volume peaked at 369 records in 2022 and had eased to 289 by 2024, an 18% decline over that three-year span — the most recent period that can be treated as a complete trend. Figures for 2025 and 2026 are still understated because publication typically lags filing by around 18 months, so it would be premature to call this a sustained slowdown based on the latest partial years. The broader pattern across 2017-2024 shows sustained, if uneven, activity rather than a single clear growth or decline trajectory.

The smaller IPC classes in this dataset — including AI-based trajectory computing, additive-manufacturing-integrated actuator control, and image-data-driven path correction — carry far fewer records than the core manipulator and control-system classes. That does not guarantee an area is unclaimed, but lower filing density in an adjacent, technically relevant branch is a reasonable starting point for a freedom-to-operate or first-claim search. Combining a white space search with a review of the strongest existing claims in the neighbouring dense classes gives a more reliable read than density figures alone.

Co-assignee filings are uncommon in this dataset: only ten co-assignee pairs appear across 4,624 records. The strongest of those pairs links a parent company to its own subsidiary rather than to an outside partner, which suggests most of what looks like collaboration is really intra-group IP structuring. External joint filings between unrelated companies are the exception rather than the norm in this field.

A representative example in this dataset, filed by KUKA Roboter GmbH, claims an industrial robot together with a path-planning method for controlling its movement, where an effector such as a remote laser welding device is mounted on the robot arm and positioned at a defined distance from a reference point on the robot. This is typical of claims sitting inside the B25J and G05B overlap: a mechanical manipulator claim combined with a motion-control method claim. Reviewing how such combined mechanism-plus-method claims are structured is useful context before drafting in the same dense claim space.