Machine Learning for Grid Operations Patents: Who Leads, Gaps 2026
- Small but concentrated field. 17 records in scope, with the top 5 assignees accounting for 64.7% of all filings.
- 2025 was the peak filing year so far. 10 of the 17 records published that year, with 2026 already showing 3 more before the data cut-off.
- Filing is split across academic and commercial actors. university and research-foundation applicants sit alongside single-inventor filers and one large platform play, with no single assignee dominating the whole field.
Top-5 share is the combined record count of the five largest assignees divided by all 17 records in scope (CR5), not by the ranked leaders only.
What this landscape covers
This review covers patent filings at the intersection of machine learning methods and grid operations — records that combine operational forecasting or security assessment language with the practical obstacles to deploying ML on live grid infrastructure: training data scarcity, physics-informed modelling, generalization to unseen conditions, explainability requirements, operator trust and deployment monitoring. It is a narrow, deliberately filtered slice of a much larger machine-learning patent universe, built to surface filings that address deployment problems rather than pure algorithmic novelty.
The scope spans 2015 through the 2026-07-31 data cut-off, with 17 total published records in the assignee ranking. Because publication typically lags filing by roughly 18 months, the 2026 count is necessarily partial and the true 2025-2026 filing volume is understated in the trend line.
Filing trend and technology composition
Two views of the same 17-record dataset: activity over time, and which IPC subclasses the filings actually sit in.
Filing trend, 2017-2026
Filings were at zero in 2017 and climbed unevenly to a peak of 10 records in 2025, the last complete year in scope. Three records have already published in 2026 ahead of the mid-year cut-off, but the true 2025-2026 total will read higher once publication lag closes.
IPC subclass composition
G06N (AI-model computing) appears on 70.6% of the 17 records, confirming this is fundamentally an AI-methods field rather than a grid-hardware one. G06F (digital data processing) reaches 41.2%, H04L (digital transmission) 23.5%, and G05B (control and regulating systems) — the subclass closest to direct grid control — sits at only 17.6%. A record can carry more than one class, so these shares add up to well over 100%.
Shares are the percentage of the 17 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Machine Learning for Grid Operations with Eureka
This page is one run against one query. Ask Eureka your own question about machine learning for grid operations and every answer comes back with the patent numbers behind it.
Try EurekaMost-cited and most recent filings
Know your model and know your data systems and methods for transactions
The filing describes configuring and deploying AI-driven transacting agents that autonomously execute transactions on behalf of an individual or organization, including granting the agent access to a digital wallet and deploying it to a public network so it can transact via digital marketplaces using predefined system prompts and configuration instructions.Filed by Strong Force TX Portfolio 2018, LLC; published 2026-01-29. Abstract condensed from the original filing.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20200184308A1 | Methods, systems, and computer readable mediums for determining a system state of a power system using a conv… | 38 |
| 2 | WO2026024864A1 | Know your model and know your data systems and methods for transactions | 12 |
Citation counts accumulate over time and favour older filings; treat them as a signal of influence within this searched corpus, not of current technical importance.
Publication numbers are shown where the record carries one (2 of 2 rows); 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 say
Read together, the concentration figures, filing trend and citation pattern point to a field that is active but not yet settled around a dominant technical approach.
A short leaderboard, not a monopoly
The top 5 assignees combined account for 64.7% of all 17 records in scope, and the top 10 reach 94.1%. With a leader holding only 3 records, no single organisation controls the field outright — but the long tail beyond the ranked leaders is thin.
2025 is the high point, 2026 already active
Filing volume rose from zero in 2017 to 10 records in 2025, the last complete year in the window. Three records already appear in 2026 ahead of the mid-year cut-off, and publication lag means this partial year understates true filing activity.
AI-methods classification dominates, direct grid control is thin
G06N appears on 70.6% of the 17 records, far ahead of G05B (control and regulating systems) at 17.6%. That gap suggests most filings are claiming the modelling and inference layer rather than the control-loop integration layer.
India and PCT filings lead over direct US filing
Among the receiving offices tracked, India accounts for 8 records and WIPO/PCT filings for 6, ahead of 3 direct United States filings. That mix points to a filing base weighted toward academic and PCT-route applicants rather than US-only commercial strategies.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to machine learning for grid operations, with the prior art for and against each one.
Who is filing, and what is still open
The ranked assignees mix academic institutions, a research foundation, individual co-inventors and one large commercial platform filer — a spread consistent with a field still being explored rather than consolidated.
A modest lead, not a lock
The top-ranked assignee holds 3 of the 17 records in scope, a lead but not a dominant one; fifth place already sits at 2 records, showing filings cluster tightly at the top rather than falling off sharply.
One filer is still active while others have gone quiet
ONMED LLC filed 2 records in the latest year, while several other ranked assignees, including Vellore Institute of Technology and Manipal University Jaipur, show 0 filings in the latest year and a -100% year-on-year change from their prior activity.
Co-filing is essentially absent
Only one co-assignee pairing appears across the 17 records, between two individual named inventors. That scarcity suggests most work in this space is being filed by single institutions or individuals rather than through joint ventures or cross-licensing arrangements.
| Assignee | Recent year | YoY |
|---|---|---|
| ONMED LLC | 2 | — |
| VELLORE INSITUTE OF TECH | 0 | -100% |
| Arris Enterprises LLC | 0 | — |
| University of Tennessee Research Foundation | 0 | — |
| MANIPAL UNIVERSITY JAIPUR | 0 | -100% |
| Telefonaktiebolaget LM Ericsson (publ) | 0 | -100% |
| Vellore Institute of Technology | 0 | -100% |
| TNTRA INC | 0 | -100% |
Where to take this
The numbers above describe the field as filed; what a specific team should do with them depends on where they sit in it.
Check freedom-to-operate before drafting
With G06N claims covering 70.6% of records and one highly-cited convolutional-network filing on system-state determination, a novelty and freedom-to-operate check against the highest-cited records is worth doing before drafting new claims in the forecasting or state-estimation space.
Run a freedom-to-operate search in EurekaTrack the quiet filers for re-entry
Several ranked assignees show 0 filings in the latest year after prior activity; renewed filing from any of them would be an early signal worth monitoring rather than assuming the space is settled.
Set up assignee monitoring in EurekaExplore the under-claimed control-integration layer
G05B (control and regulating systems) sits at only 17.6% of records despite G06N modelling claims reaching 70.6%, suggesting the control-loop integration of ML outputs is comparatively open.
Explore white space in EurekaCommon questions
The dataset ranks 12 companies and institutions, with the top-ranked assignee holding 3 of the 17 records in scope and the top 5 combined accounting for 64.7% of all records. The field mixes academic institutions, a research foundation, individually named inventors and one large commercial platform filer, so leadership is spread thinly rather than concentrated in one organisation. Anyone assessing competitive risk should look at the full ranked list rather than assume a single dominant player, since even the leader holds only a small fraction of total filings.
This landscape identifies 17 published records matching the search scope, covering filings from 2015 through the 2026-07-31 data cut-off. That is a narrow, filtered count built around deployment-specific language such as training data scarcity, explainability and operator trust, not a count of all machine-learning-for-grids patents broadly. Filing peaked at 10 records in 2025, and 2026 already shows 3 more before the cut-off, though publication lag means recent-year counts will rise as more filings publish.
G06N, covering AI-model computing, appears on 70.6% of the 17 records, making it by far the dominant classification. G06F (digital data processing) follows at 41.2%, H04L (digital transmission) at 23.5%, and G05B (control and regulating systems) at only 17.6%. Because a single record can carry multiple IPC classes, these percentages add up to well over 100%, and the low G05B share suggests most filings claim the modelling layer rather than direct grid control integration.
Filing rose from zero recorded activity in 2017 to a peak of 10 records in 2025, the most recent complete year in the dataset, with 3 more records already published in 2026 ahead of the mid-year cut-off. A precise growth rate is not calculable here because fewer than four complete years remain once the roughly 18-month publication lag is accounted for. The clear takeaway is that 2025 was the busiest year on record, but the true scale of 2025-2026 activity will only become visible as later filings publish.
The clearest gap sits between the heavily-claimed AI-modelling layer (G06N at 70.6% of records) and the much thinner control-integration layer (G05B at only 17.6%), pointing to open space in closing the loop between ML forecasts and live grid control actions. Deployment monitoring, explainability outputs aimed at operator sign-off, and physics-informed generalization to unseen grid topologies are also represented in only a handful of the 17 records. Co-filing is nearly absent too, with only one identified co-assignee pairing, suggesting joint development and licensing structures in this space remain largely untried.
Research Machine Learning for Grid Operations in depth with Eureka
Go past this page: query the whole machine learning for grid operations 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.