Humanoid Robots Patents: Who Leads, Where the Gaps Are 2026
- Concentrated but not closed. The top 5 assignees hold 26.8% of all 2,085 records in scope, and the top 10 hold 37.6% — leaving most of the field to a long tail of single- and few-filing entrants.
- Growth is real, not a spike. Filings rose from 114 in 2021 to 135 in 2024, an 18% increase over the three-year span that publication lag hasn't yet caught up to.
- Claim space is lopsided. B25J (manipulators & robots) covers 59.2% of records, while perception classes like G06V sit at just 3.9% — control-side claims are dense, perception-side claims are comparatively open.
Filing growth compares 2021 (114 records) with 2024 (135) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 2,085 records in scope (CR5), not by the ranked leaders only.
What the humanoid robot patent record actually shows
The search set covers 2,085 published records filed against humanoid robot mechanisms and control together with position estimation, trajectory, sensing, actuator control or path planning — narrowing the field to robots that move and sense, not every robotics filing that happens to mention a humanoid form. Filing activity has grown steadily since 2017, reaching a peak so far in 2025, though the most recent one to two years are understated because publication lags filing by roughly 18 months.
Ownership is concentrated at the top around a small set of automotive, consumer-electronics and dedicated robotics firms, but the ranked leaders account for well under half of all records — most of the corpus sits with entities that filed only a handful of times. Technology composition skews heavily toward mechanical manipulation and motion control, with computing, AI and vision classes present but comparatively thinner.
Filing trends and technology composition
Two views of the same 2,085-record corpus: filings by year, and how records distribute across IPC subclasses. Because a single record can carry several IPC codes, the class shares below add up to more than 100% of records — that's expected, not an error.
Filings climbed through the decade, with 2025 as the peak year so far
Annual filings moved from 66 in 2017 to a peak of 235 in 2025. The 2021→2024 span alone shows +18% growth (114 to 135), a complete-year comparison that avoids the understated tail of 2025-2026 filings still working through publication.
Manipulator and control classes dominate; vision and AI trail
B25J (manipulators & robots) appears on 59.2% of the 2,085 records, far ahead of G06F electric data processing (16.2%), G05D non-electric variable control (14.8%) and G05B control/regulating systems (12.0%). AI-specific computing (G06N, 11.2%) and vehicle steering crossover (B62D, 11.1%) trail those, while image processing (G06T, 6.8%) and recognition (G06V, 3.9%) are the thinnest represented classes despite being central to real-world deployment.
Shares are the percentage of the 2,085 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Humanoid Robots Patent Landscape with Eureka
This page is one run against one query. Ask Eureka your own question about humanoid robots patent landscape and every answer comes back with the patent numbers behind it.
Try EurekaA representative claim and the most-cited prior art
US20240181637A1 — Autonomous humanoid robot
An autonomous humanoid robot configured to overcome limited maneuvering issues by offering a more efficient autonomous humanoid robot that autonomously operates to interact with users and interact with other robots, and includes a computing system configured to provide instruction and programming for estimating and controlling pivotal movement of body components involving arms, legs and a waist module which are configured to support the body and reposition the body such that the autonomous humanoid robot can step, walk, roll or skate or perform various handling maneuvers to complete tasks.Filed by an individual assignee (Gillett, Carla R.), published 2024-06-06 — illustrating that meaningful claim territory in this field is not exclusively held by large corporate filers.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20220066456A1 | Obstacle recognition method for autonomous robots | 763 |
| 2 | US20210089040A1 | Obstacle recognition method for autonomous robots | 458 |
| 3 | US20150290795A1 | Methods and systems for food preparation in a robotic cooking kitchen | 457 |
| 4 | US20160059412A1 | Robotic manipulation methods and systems for executing a domain-specific application in an instrumented envir… | 452 |
| 5 | US20090055019A1 | Interactive systems employing robotic companions | 419 |
| 6 | US20190291277A1 | Systems and methods for operating a robotic system and executing robotic interactions | 373 |
| 7 | US20090081923A1 | Robotic game systems and methods | 365 |
| 8 | US20070016329A1 | Biomimetic motion and balance controllers for use in prosthetics, orthotics and robotics | 338 |
| 9 | US20110071675A1 | Visual perception system and method for a humanoid robot | 300 |
| 10 | US20070255454A1 | Control Of Robots From Human Motion Descriptors | 278 |
Citation counts reflect influence within this searched corpus and skew toward older filings — a highly cited 2015 record is not necessarily more relevant today than a lightly cited 2023 one.
Each row carries its publication number; clicking a row searches Eureka by that number.
Put your own technology through the same analysis
Eureka on the web
When you want the answer in the next five minutes.
The agent works the prompt against patents and technical literature, citing every source.
Run your analysis now →MCP server & REST API
When it has to run inside your own pipeline.
Patent search, landscape analysis and assignee resolution as MCP tools. Drop them into any agent framework, or call REST directly.
Browse MCP servers →What the numbers mean for filing strategy
Three signals worth acting on: where claim density already blocks easy entry, where growth is durable rather than a one-year spike, and where citation influence is concentrated in older art.
Leadership is real but not exclusionary
The top 5 assignees combine for 26.8% of all records and the top 10 for 37.6% — a meaningful cluster, but it leaves the majority of the corpus to firms with a handful of filings each. New entrants are competing against a moderately dense leadership tier, not a closed field.
Growth is a multi-year trend, not a 2025 blip
Filings rose from 114 in 2021 to 135 in 2024 — the last span both years of which are complete enough to compare reliably. The apparent 2025 peak and 2026 partial-year figures should be read as still filling in, not as a plateau.
Mechanical claims are dense; vision claims are thin
B25J manipulator and robot claims sit on well over half the corpus, while image/video recognition (G06V) appears on under 4%. That gap suggests perception and recognition claims tied to humanoid platforms remain comparatively open relative to mechanism and actuation claims.
Influence skews toward earlier obstacle-recognition art
The most-cited records in this corpus concern obstacle recognition and general-purpose robotic manipulation, several dating to the mid-2010s. High citation counts here mark historical influence within the searched set, not necessarily current commercial relevance.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to humanoid robots patent landscape, with the prior art for and against each one.
Who is filing, and where momentum is shifting
The ranked assignee list spans automotive OEMs, consumer electronics firms and dedicated robotics companies. Recent-year momentum figures should be read cautiously: the latest year is partial, so year-over-year drops for several leaders reflect the publication lag as much as any real pullback.
A single leader sits well ahead of fifth place
The top-ranked assignee holds 214 records against 56 for fifth place and 40 for tenth — a steep drop-off that marks a genuine leadership gap rather than a tightly bunched top tier.
Recent-year drops likely reflect publication lag, not retreat
Several major filers show sharp year-over-year declines in the latest year, including drops of -81%, -80% and -100% among named leaders. Given the roughly 18-month gap between filing and publication, these should not be read as filers exiting the space.
Filing is US-centred with a strong PCT and EPO presence
The United States receives the largest share of filings at 1,031, ahead of Europe (295), China (201), WIPO/PCT (172), Germany (67) and India (65). That spread points to a field being protected across major jurisdictions rather than concentrated in one home market.
| Assignee | Recent year | YoY |
|---|---|---|
| NVIDIA Corporation | 5 | -81% |
| Boston Dynamics Inc. | 2 | -33% |
| UBTECH Robotics Corp Ltd | 1 | -80% |
| Honda Motor Co., Ltd. | 0 | -100% |
| Sony Group Corporation | 0 | — |
| Samsung Electronics Co., Ltd. (Korea) | 0 | -100% |
| Honda Research Institute Europe | 0 | — |
| MBL Ltd | 0 | — |
Where to take this next
The dataset points to specific next steps depending on whether the goal is freedom-to-operate, whitespace scouting or watching a competitor.
Run a freedom-to-operate check on B25J claims
With 59.2% of records carrying a B25J manipulator/robot code, any new mechanism design should be checked against this class first, where claim density is highest.
Explore claim scope in EurekaScout the vision and AI-model white space
G06V and G06N classes carry far fewer records than B25J or G06F, suggesting recognition- and AI-model-driven claims tied to humanoid platforms are less crowded.
Map white space in EurekaTrack the leadership tier's actual filing cadence
Momentum figures for the latest year are distorted by publication lag; a proper competitor watch needs a rolling view that accounts for the 18-month gap.
Set up monitoring in EurekaCommon questions about humanoid robot patents
The ranked assignee list is led by a single company with 214 records, well ahead of the fifth-ranked firm at 56 and the tenth at 40. The leader tier includes automotive OEMs, consumer electronics firms and dedicated robotics companies, but the top 10 combined only account for 37.6% of the 2,085 records in scope, so leadership is real but far from a monopoly. Anyone assessing competitive position should look at the full ranked list rather than assume the field is a two- or three-way race.
Filings grew from 114 in 2021 to 135 in 2024, an 18% increase over that three-year span, which is the most reliable complete-year comparison available. Figures for 2025 and 2026 appear lower in raw counts, but that reflects the roughly 18-month lag between filing and publication, not a real slowdown. Treat any apparent recent-year dip as provisional until later publication catches up.
Mechanical manipulation and robot structure (IPC class B25J) dominate, appearing on 59.2% of the 2,085 records in scope. Control-related classes — electric data processing (G06F), non-electric variable control (G05D) and control/regulating systems (G05B) — each cover between 12% and 16% of records. Vision and recognition classes (G06T, G06V) are comparatively thin, at 6.8% and 3.9% respectively, despite their importance to real-world navigation and interaction.
The clearest gap sits between mechanism-heavy claims, which are dense, and perception- or AI-model-driven claims, which are comparatively sparse: G06V image/video recognition covers under 4% of records against B25J's nearly 60%. Sub-areas like visual scene recognition for bipedal navigation, image-based grasp planning and multi-sensor balance fusion look under-claimed relative to their practical role in deployed systems. That said, thin filing density in a subclass is not proof the area is legally open — it means fewer close prior-art hits, and any filing there still needs its own search.
Moderately concentrated at the top and long-tailed beneath it: the top 5 assignees hold 26.8% of the 2,085 records in scope, and the top 10 hold 37.6%. That leaves roughly two-thirds of the corpus spread across the remaining ranked companies and unranked single-filers. It is a field with an identifiable leadership tier but plenty of room for new entrants to stake out claims outside that tier's core areas.
Research Humanoid Robots Patent Landscape in depth with Eureka
Go past this page: query the whole humanoid robots patent landscape corpus yourself, in your own scope.
Every answer comes back with patent numbers you can open.
Disclaimer. This page is generated from Patsnap Eureka data drawn from a limited snapshot of global patent and scientific-literature records, and is provided for general information and reference only.
Patent data carries inherent limitations: recent filings (typically the most recent 18–24 months) are under-counted due to standard publication lag; counts may be reported at either a patent-family or a patent-record basis and are not always directly comparable; classification, applicant-name, and citation data may contain errors, duplicates, or omissions; and the underlying search query defines and constrains the scope shown. As a result, the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing on this page constitutes an exhaustive prior-art, novelty, freedom-to-operate, or validity search, nor does it constitute legal, financial, investment, or professional advice, and it should not be relied upon as such. Any patent, commercial, or strategic decision should be verified independently and reviewed with qualified patent, legal, and domain professionals. Patsnap makes no warranties, express or implied, as to the accuracy, completeness, or fitness for any particular purpose of the information presented.
Machine translation. Assignee and organisation names originally recorded in Chinese, Japanese or Korean have been rendered into English by an AI translation step so that the tables stay readable. These renderings are best-effort and may not match a company’s registered English name; the original name is what the underlying patent record carries, and it is what any Eureka query launched from this page uses.