Knowledge Graph Patents: Who Leads, Where the Gaps Are 2026
- 22.7% of all 4,245 records sit with just five assignees, while the ranked leaders' list runs 100 deep with a long tail of single- and few-filing entrants.
- Filings rose 47% from 435 in 2021 to 640 in 2024, the last year publication lag lets us treat as complete.
- G06F digital data processing covers 68.2% of records, but healthcare informatics (G16H, 5.2%) and control systems (G05B, 5.0%) remain comparatively open.
Filing growth compares 2021 (435 records) with 2024 (640) — 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 4,245 records in scope (CR5), not by the ranked leaders only.
What this landscape covers
This review covers 4,245 published records matching graph-database and knowledge-graph search terms combined with data-infrastructure terms such as query indexing, schema mapping, and record storage — filed between 2015 and the 2026-08-31 cut-off. Filing activity is concentrated in general digital data processing, but a meaningful share of records also touch AI computing methods and network transmission, reflecting how knowledge graphs are increasingly claimed as infrastructure underneath AI and networking systems rather than as a standalone data structure.
Because publication typically lags filing by around 18 months, the most recent one to two years in any trend understate real filing activity; treat 2024 as the last complete year of the reported growth rate.
Filing trends and technology composition
Filing volumes and IPC class distribution for the 4,245 records in scope, drawn directly from the underlying dataset.
Filing trend, 2017–2026
Annual publications rose from 157 in 2017 to a peak of 640 in 2024, up 47% from 2021's 435. 2025 and 2026 figures are still filling in as later-filed applications publish.
IPC subclass composition
G06F (electric digital data processing) appears in 68.2% of records, followed by G06N (AI-based computing, 29.6%) and H04L (digital information transmission, 22.5%). Records commonly carry multiple IPC classes, so these shares sum to well over 100%.
Shares are the percentage of the 4,245 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Knowledge Graphs & Graph Computing Patent Landscape with Eureka
This page is one run against one query. Ask Eureka your own question about knowledge graphs & graph computing patent landscape and every answer comes back with the patent numbers behind it.
Try EurekaRepresentative and most-cited filings
EP4791059A1 — Managing a mobile communication network using a semantic knowledge graph
The application describes modelling a mobile communication network and its telecommunication domain knowledge as a semantic knowledge graph — capturing structured concepts, taxonomies, relationships, and constraints — then obtaining runtime data from the network and correlating it against that graph to enable logical inference over network state.Filed by Deutsche Telekom, published 2026-08-12, this is one of the most recent applications in the dataset and illustrates the shift from generic graph-storage claims toward domain-specific ontology modelling for network operations.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20160012465A1 | System and method for distributing, receiving, and using funds or credits and apparatus thereof | 2,559 |
| 2 | US9516053B1 | Network security threat detection by user/user-entity behavioral analysis | 1,294 |
| 3 | US20190339688A1 | Methods and systems for data collection, learning, and streaming of machine signals for analytics and mainten… | 974 |
| 4 | US20200348662A1 | Platform for facilitating development of intelligence in an industrial internet of things system | 732 |
| 5 | US20210157312A1 | Intelligent vibration digital twin systems and methods for industrial environments | 715 |
| 6 | US20210090694A1 | Data based cancer research and treatment systems and methods | 714 |
| 7 | US20190042988A1 | Omnichannel, intelligent, proactive virtual agent | 632 |
| 8 | US20200103894A1 | Methods and systems for data collection, learning, and streaming of machine signals for computerized maintena… | 514 |
| 9 | US10673880B1 | Anomaly detection to identify security threats | 453 |
| 10 | US20180219888A1 | Graph-Based Network Security Threat Detection Across Time and Entities | 452 |
Citation counts favour older filings that have had more time to accumulate citations within the searched corpus; read them as a signal of influence, not current relevance.
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 concentration and growth figures mean
Three patterns stand out once filing volume, class composition and citation depth are read together.
Claim density sits with a small group, but the tail is long
The top 5 assignees combined account for 22.7% of all 4,245 records, and the top 10 combine for 32.4%. That leaves roughly two-thirds of filings spread across a ranked list of 100 companies, many with only a handful of families each — a sign that core graph-storage and query-indexing claims are contested, but adjacent application areas remain open to newer entrants.
Filing activity accelerated into 2024
Annual publications grew from 435 in 2021 to 640 in 2024, a 47% rise over that span. Because publication lags filing by around 18 months, 2025 and 2026 volumes will continue to rise as later applications publish — the apparent tapering in the two most recent years is a reporting artefact, not a slowdown.
Most claims sit in general data processing, not AI methods specifically
G06F (electric digital data processing) appears in 68.2% of records, well ahead of G06N (AI-based computing, 29.6%) and H04L (digital transmission, 22.5%). Smaller classes — G16H healthcare informatics at 5.2% and G05B control systems at 5.0% — show where graph techniques are being applied to a specific vertical rather than claimed as general infrastructure.
US filing dominates, PCT and EPO follow at a distance
United States filings (2,755) outnumber the next largest venues, WIPO/PCT (447) and the EPO (411), by a wide margin. India (157), Canada (121) and Australia (86) trail further behind, suggesting most applicants are still treating the US as the primary market before deciding on wider international protection.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to knowledge graphs & graph computing patent landscape, with the prior art for and against each one.
Who is filing, and where the gaps sit
The assignee ranking spans 100 companies from large technology platforms to portfolio-holding entities; momentum in the most recent year has slowed across nearly every top filer, consistent with publication lag rather than a genuine pullback.
One filer holds a clear lead
The top-ranked assignee holds 313 records, well ahead of the fifth-placed filer at 110 and the tenth-placed filer at 71 — a steep drop-off that indicates a genuine leader rather than a cluster of near-equal filers at the top.
Recent-year counts are down across the board
Several of the most active assignees show sharp year-over-year declines in the latest year, including drops of -80% and -100%. Given the roughly 18-month publication lag, this reflects incomplete recent-year data rather than an actual retreat from the space.
Co-filing is rare and concentrated in one corporate family
Only three co-assignee pairs appear in the dataset, and all three involve the same corporate group filing jointly across its national subsidiaries. Cross-company co-filing is essentially absent, which suggests graph-computing IP here is developed and held internally rather than through joint ventures.
| Assignee | Recent year | YoY |
|---|---|---|
| Wiz Inc | 9 | -88% |
| STRONG FORCE TX PORTFOLIO 2018 LLC | 2 | -80% |
| Oracle International Corporation | 2 | -80% |
| Strong Force IoT Portfolio 2016 LLC | 1 | -67% |
| Strong Force VCN Portfolio 2019 LLC | 0 | -100% |
| Splunk Inc | 0 | — |
| Microsoft Technology Licensing, LLC | 0 | -100% |
| International Business Machines Corporation (IBM) | 0 | -100% |
Where to take this analysis
The figures above establish where claim density sits today; the next step is usually narrower, mapping specific claim language against a candidate filing or freedom-to-operate question.
Check a draft claim against this landscape
Run a specific claim concept through the dense G06F and G06N territory identified here to see which granted claims it would need to design around.
Explore in EurekaTrack the leading assignees' recent activity
Follow filing behaviour from the top-ranked assignees identified in this dataset as their 2025–2026 applications continue to publish.
Explore in EurekaInvestigate the under-claimed branches
Healthcare informatics and control-system applications of graph techniques show thinner coverage than the core infrastructure claims — worth a closer freedom-to-operate check before committing R&D.
Explore in EurekaCommon questions about this landscape
One assignee leads the ranked list with 313 records, noticeably ahead of the fifth-placed filer at 110 and the tenth-placed filer at 71. The top 5 assignees combined account for 22.7% of all 4,245 records in scope, and the top 10 combined account for 32.4%. Beyond that group, coverage spreads thinly across a ranked list of 100 companies, many holding only a handful of families, so the field is led by a small group but not monopolised by it.
Yes, based on the last fully reported period: annual publications rose from 435 in 2021 to 640 in 2024, a 47% increase over three years. Filing counts for 2025 and 2026 appear lower in the raw data, but that reflects the roughly 18-month lag between filing and publication rather than an actual slowdown, since later applications from those years are still working through the pipeline. 2024 is the most recent year that can be read as a complete picture.
The largest IPC category is G06F, electric digital data processing, appearing in 68.2% of the 4,245 records in scope. AI-based computing (G06N) follows at 29.6%, and digital information transmission (H04L) at 22.5%. Smaller categories such as healthcare informatics (G16H, 5.2%) and control and regulating systems (G05B, 5.0%) show where graph techniques are applied to a specific industry vertical rather than claimed as general-purpose infrastructure, since a single record can carry multiple IPC classes these figures add to more than 100%.
The core storage, indexing and query claims in G06F are the most densely filed and hold the concentration described above, making them the hardest area to stake new ground in. Comparatively thinner classes such as healthcare informatics, control-system integration and wireless-network applications of graph techniques show less claim density, which is generally where first-mover claims are easier to draft without running into a crowded prior-art field. Any specific filing decision should still be checked against the granted claims of the leading assignees named in the ranking, not just the class-level counts.
EP4791059A1, filed by Deutsche Telekom and published 2026-08-12, describes managing a mobile communication network by modelling the network and its telecom domain knowledge as a semantic knowledge graph — including structured concepts, taxonomies, relationships and constraints — and then correlating live runtime network data against that graph to support inference. It is one of the most recent filings in this dataset and reflects a broader trend toward domain-specific ontology modelling rather than generic graph-database claims. Anyone building similar network-management inference tools should review its specific claim scope directly rather than relying on this summary alone.
Research Knowledge Graphs & Graph Computing Patent Landscape in depth with Eureka
Go past this page: query the whole knowledge graphs & graph computing 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.