Fastest
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 →https://www.patsnap.com/resources/blog/rd-blog/rram-device-variability-for-analog-compute-patent-landscape/ · Patsnap · data cut-off 2026-07-31 · downloaded from the live page
Patent Landscape · Memory & StorageRRAM Device Variability for Analog Compute Patents
- 33.5% of the field sits with five companies. The top 5 assignees account for 1,084 of 3,239 records in scope, while a long tail of single- and low-filing entrants fills out the rest of the ranked leaders.
- Filings have cooled since a 2019 peak of 271. The most recent full year sits well below that high point, and the partial current-year count of 10 understates real activity given the usual 18-month publication lag.
- Claim space is concentrated in two IPC subclasses. H01L covers 62.0% of records and G11C 38.3%, meaning device structure and digital memory architecture, not analog-specific programming, still absorb most of the drafting.
3,239
Published Records
33%
Top-5 Share of All Records
-42%
Filing Growth 2021→2024
US
Leading Jurisdiction
Filing growth compares 2021 (220 records) with 2024 (128) — 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 3,239 records in scope (CR5), not by the ranked leaders only.
Published byPatsnap Research··6 min readSourced from Patsnap Eureka
Overview
What this landscape covers
This landscape tracks patent filings addressing RRAM device variability as it bears on analog compute: conductance drift, write-verify programming, retention behaviour and the mapping errors that arise when a resistive cell is asked to hold an analog weight rather than a binary state. The scope spans 3,239 published records filed between 2015 and mid-2026, drawn from filings that combine RRAM or analog conductance programming language with claim terms around cycling, device-level programming and retention.
The dataset sits at the intersection of memory fabrication and compute-in-memory architecture, so a single record commonly carries classes for both the physical device and the digital memory system built on top of it. That overlap is visible in the technology composition below, and it is one reason class shares in this field add up to well over 100% of the record total.
Filing activity and technology composition, 2015-2026
Source: Patsnap Eureka. Assignee ranking and totals. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Let an AI agent run this analysis on your own technology
Pick a task. Every answer cites the patents behind it.
10,000 free credits to start
The Data
Filing trend and technology composition
Two views of the same 3,239 records: how filing activity has moved year over year, and which IPC subclasses carry the claim volume.
Filing trend, 2017-2026
Filings ran at 207 in 2017, rose to a peak of 271 in 2019, and had fallen to a midpoint of 153 by 2022. The count of 10 for the most recent year is partial and should be read against the 18-month gap between filing and publication, not as a cliff in real activity.
IPC subclass concentration
H01L (semiconductor devices) appears on 62.0% of records and H10N (other electric solid-state devices) on 24.5%, alongside G11C (static and digital memories) at 38.3% and H10B (memory device manufacture) at 20.4%. Because records carry multiple classes, these figures describe overlapping coverage, not a partition of the field.
Shares are the percentage of the 3,239 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Source: Patsnap Eureka. Filing trend and technology composition. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.
Go deeper on RRAM Device Variability for Analog Compute with Eureka
This page is one run against one query. Ask Eureka your own question about rram device variability for analog compute and every answer comes back with the patent numbers behind it.
Try EurekaKey Patents
Most-cited records in this landscape
Representative filing
Test device and test method for resistive random access memory and resistive random access memory device
KIOXIA CORPORATION
The filing describes a test arrangement that drives core control signals from two write-enable signals with different timing behaviour, letting an operator verify RRAM operation by locally shortening the interval between core control signal transitions relative to the write-enable period.US20120079330A1 — filed by Kioxia Corporation, 2012-03-29
Highest-citation records, 2015-2026 scope
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20140322862A1 | Method of making a resistive random access memory device with metal-doped resistive switching layer | 434 |
| 2 | US20130288427A1 | Methods Of Fabricating Dielectric Films From Metal Amidinate Precursors | 429 |
| 3 | US20140322885A1 | Method of making a resistive random access memory device | 427 |
| 4 | US9142764B1 | Methods of forming embedded resistors for resistive random access memory cells | 349 |
| 5 | US20190198571A1 | Resistive memory with a plurality of resistive random access memory cells each comprising a transistor and a … | 334 |
| 6 | US20100112810A1 | Resistive random access memory and method for manufacturing the same | 261 |
| 7 | US20150034898A1 | Confined Defect Profiling within Resistive Random Memory Access Cells | 170 |
| 8 | US20090272962A1 | Reduction of forming voltage in semiconductor devices | 165 |
| 9 | US20070257300A1 | Structures and Methods of a Bistable Resistive Random Access Memory | 162 |
| 10 | US6982902B2 | MRAM array having a segmented bit line | 150 |
Citation counts favour older filings simply because they have had longer to accumulate citations inside this corpus; treat the ranking as a signal of influence on later filers, not of current commercial relevance.
Each row carries its publication number; clicking a row searches Eureka by that number.
Source: Patsnap Eureka. Citation counts and representative records. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Run it yourself
Put your own technology through the same analysis
Where to run it
For builders
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 →Insights
What the numbers mean for a filing decision
Three patterns from the dataset that matter more than the raw counts on their own.
Concentration
33.5%
of 3,239 records held by top 5
The top of the field is dense, the tail is not
A third of all records in scope trace to five assignees, and the tenth-ranked filer alone still holds 77 records. Below that, the ranked leaders list thins quickly into single-digit filers, which is where freedom-to-operate searches usually turn up surprises rather than blockers.
Top 10 combined: 46.8% of records
Momentum
271 → 10
peak year (2019) vs latest partial year
Activity has cooled from its 2019 peak
Filing volume has trended down since 2019, and several previously active assignees show 0% or negative year-on-year change in the most recent period. Read the latest year's low count with the 18-month publication lag in mind rather than as evidence the field has stopped.
Midpoint year 2022: 153 filings
Composition
62.0%
of records classed H01L
Device structure still dominates over programming method
H01L and H10B together point to a field still weighted toward fabrication and device structure claims, while G06F at 4.5% shows comparatively little claim volume has moved up the stack into the digital processing side of analog compute.
G06F (digital data processing): 4.5% of records
Eureka AI Agent
Looking for what nobody has claimed yet?
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to rram device variability for analog compute, with the prior art for and against each one.
Source: Patsnap Eureka. Co-assignee relationships and derived observations. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Players
Who is filing, and where the field still opens
The ranking below covers 100 companies counted by patent family; it is the full set the data endpoint returns, not a curated top list.
Leader
423
records, leading assignee
One filer sits well ahead of the field
The leading assignee holds 423 records against a fifth-place figure of 103 and a tenth-place figure of 77 — a steep drop-off that marks this as a leader-plus-tail structure rather than a tightly clustered top group.
Fifth place: 103 records
Co-filing
47
shared records, strongest pair
A small set of co-assignee relationships repeats
Only 10 co-assignee pairs appear in the dataset, and the strongest recurring pairings sit among a small cluster of memory specialists, suggesting joint development work is concentrated rather than widespread across the field.
10 co-assignee pairs identified
Momentum
-100%
YoY, several prior filers
Recent-year activity has gone quiet for past leaders
Multiple assignees that filed steadily in earlier years show 0 records in the latest year and year-on-year drops as steep as -100%, consistent with the broader slowdown from the 2019 peak rather than any single company's exit.
One assignee still filed in the latest year, though down 93% YoY
Recent-year filing momentum by assignee
| Assignee | Recent year | YoY |
|---|---|---|
| Taiwan Semiconductor Manufacturing Co., Ltd. (TSMC) | 1 | -93% |
| United Microelectronics Corp (UMC) | 1 | 0% |
| International Business Machines Corporation (IBM) | 0 | -100% |
| Winbond Electronics Corp | 0 | -100% |
| Intermolecular Inc | 0 | — |
| Samsung Electronics Co., Ltd. | 0 | -100% |
| Institute of Microelectronics, Chinese Academy of Sciences | 0 | -100% |
| Peking University | 0 | -100% |
Source: Patsnap Eureka. Assignee-level momentum. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
What's Next
Where to take this from here
The landscape points to where claim space is full and where it is not; the next step is testing a specific claim or design against it.
Check freedom-to-operate on a specific mechanism
Device structure and fabrication claims are dense under H01L and H10B. A programming-method or drift-correction claim is more likely to clear than another device-structure variant.
Run a freedom-to-operate search in EurekaTrack the leader-plus-tail structure over time
With filings down from their 2019 peak and several past leaders quiet in the latest year, watching which assignees resume filing is more informative than the raw yearly count.
Set up assignee monitoring in EurekaSource: Patsnap Eureka. Forward-looking reading of the same dataset. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
FAQ
Common questions on this landscape
The dataset behind this landscape contains 3,239 published records filed between 2015 and mid-2026 that match RRAM or analog conductance programming language combined with claim terms around cycling, device-level programming, retention or mapping error. This is a scoped search result, not a count of every RRAM patent ever filed, since it specifically targets variability-related claim language rather than RRAM generally. Anyone repeating the count should keep the same search terms and date window to get a comparable figure.
The leading assignee holds 423 records, well ahead of the fifth-ranked filer at 103 and the tenth at 77, which marks a leader-plus-tail structure rather than a tightly bunched top group. The top 5 assignees together account for 33.5% of all 3,239 records in scope, and the top 10 for 46.8%. Below the ranked leaders the field thins into many low-volume filers, so a competitive read should weight the leading handful heavily and treat the rest as a long tail rather than equal peers.
Filing volume peaked at 271 in 2019 and had fallen to 153 by the 2022 midpoint, with the most recent year showing only a partial count of 10. Some of that decline reflects the roughly 18-month lag between filing and publication, which always understates the newest year in any patent trend. It may also reflect the field maturing past its initial wave of device-structure claims, consistent with several previously active assignees showing flat or negative year-on-year filing counts.
US20120079330A1 covers a specific test arrangement for RRAM devices: driving core control signals from two write-enable signals with different timing characteristics so that operation can be verified by checking timing margins between word-line and bit-line activation. It blocks that particular test-signal generation and verification method, not RRAM devices or write-verify programming in general. A team using a different signal-timing approach to verify RRAM operation would need to check this filing's specific claim language rather than assume it covers all test-mode verification schemes.
IPC composition shows heavy concentration in device structure and fabrication (H01L at 62.0% of records, H10B at 20.4%) but comparatively little in digital data processing (G06F at 4.5%), suggesting programming-side and system-integration claims are less crowded than device claims. Areas like write-verify convergence algorithms, cross-cell conductance mapping correction and retention-aware re-programming show thinner filing density than the core fabrication claims. These are reasonable starting points for a claim search, though thin filing density in a public search does not by itself guarantee a claim will be granted.
Answers are grounded in the same dataset. Derived from a Patsnap search on RRAM Device Variability for Analog Compute covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Research RRAM Device Variability for Analog Compute in depth with Eureka
Go past this page: query the whole rram device variability for analog compute 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.