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/in-memory-computing-patent-landscape-patent-landscape/ · Patsnap · data cut-off 2026-08-31 · downloaded from the live page
Patent Landscape · High-Performance & Advanced ComputingIn-Memory Computing Patents: Who Leads and Where Filings Are Headed
- Concentrated at the top. The five leading assignees hold 64.6% of all 3,721 records in scope, and the top ten hold 73.0% — a small group controls most of the claim space.
- Filings cooled from a 2022 peak. Activity peaked at 633 records in 2022; the last complete year shows filings down 32% from 2021's 580 to 2024's 396.
- G06F dominates, but not exclusively. 81.8% of records carry a G06F digital-data-processing class, while AI-model computing (G06N, 11.4%) and static memory (G11C, 10.2%) mark the adjacent growth edges.
3,721
Published Records
65%
Top-5 Share of All Records
-32%
Filing Growth 2021→2024
US
Leading Jurisdiction
Filing growth compares 2021 (580 records) with 2024 (396) — 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 3,721 records in scope (CR5), not by the ranked leaders only.
Published byPatsnap Research··6 min readSourced from Patsnap Eureka
Overview
What this landscape covers
In-memory computing moves computation into or next to memory arrays to cut the data-movement cost that dominates conventional von Neumann architectures. This landscape draws on 3,721 published records matched against in-memory and compute-in-memory terminology paired with system-level concerns such as cache coherence, memory access and interconnect fabric — filings that treat the memory-compute boundary as an engineering problem, not just a memory-cell chemistry claim. Publication lags filing by roughly 18 months, so the most recent one to two years in any trend understate real activity.
The scope spans processor and memory vendors, systems integrators and research consortia, filing across US, European, PCT, German and Chinese offices. Family counts, rather than raw document counts, are used for the assignee ranking so that multi-jurisdiction and continuation filing does not distort who actually holds the largest claim positions.
Filing activity and technology composition, 2017-2026
Source: Patsnap Eureka. Assignee ranking and totals. Derived from a Patsnap search on In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
The data
Filing trend and technology composition
Two views of the same 3,721-record set: filings by year, and the IPC subclasses those records carry.
Filings rose sharply through 2022, then eased
Annual filings climbed from 76 in 2017 to a peak of 633 in 2022. The most recent complete comparison — 2021's 580 against 2024's 396 — shows a 32% pullback; 2025 and 2026 figures are still incomplete due to publication lag and should not be read as a continued decline.
G06F anchors the field; AI and memory classes trail
G06F (electric digital data processing) appears on 81.8% of records, confirming that most filings frame in-memory computing as a data-processing architecture problem. H04L (digital transmission, 18.5%), G06N (AI-model computing, 11.4%) and G11C (static and digital memories, 10.2%) are the next-largest classes, each describing a different angle on the same underlying memory-compute boundary. Because records carry multiple classes, these shares are read against the full 3,721-record total, not against each other.
Shares are the percentage of the 3,721 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 In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. Counts reflect published records only and shift as new filings publish.
Go deeper on In-Memory Computing Patent Landscape with Eureka
This page is one run against one query. Ask Eureka your own question about in-memory computing patent landscape and every answer comes back with the patent numbers behind it.
Try EurekaKey patents
Most-cited records in the field
Representative filing
In memory computing processor and method thereof with direction-based processing (US20240111828A1)
SAMSUNG ELECTRONICS CO., LTD.
The filing discloses an SRAM-based in-memory computing processor using two macro types distinguished by whether the data-write direction matches the multiply-accumulate operation direction, with a shift accumulator combining their outputs for feature-map and weight computation.Assignee and filing date are rendered separately from this abstract summary.
Highest-citation records in scope
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US6807614B2 | Method and apparatus for using smart memories in computing | 510 |
| 2 | US7546438B2 | Algorithm mapping, specialized instructions and architecture features for smart memory computing | 432 |
| 3 | US4403283A | Extended memory system and method | 219 |
| 4 | US20250259085A1 | Convergent Intelligence Fabric for Multi-Domain Orchestration of Distributed Agents with Hierarchical Memory … | 209 |
| 5 | US5255211A | Methods and apparatus for generating and processing synthetic and absolute real time environments | 206 |
| 6 | US20220027051A1 | Data Path Virtualization | 186 |
| 7 | US5759044A | Methods and apparatus for generating and processing synthetic and absolute real time environments | 182 |
| 8 | US20030014200A1 | Revenue meter with power quality features | 170 |
| 9 | US20190042518A1 | Platform interface layer and protocol for accelerators | 160 |
| 10 | US20190356736A1 | Network authentication for a multi-node array | 159 |
Citation counts favour older filings that have had more time to accumulate references — treat them as a signal of influence within this corpus, 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 In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. 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 read-throughs from the concentration, class and citation data above.
Concentration
64.6% top-5 share
of 3,721 records
A small group holds most of the field
The five leading assignees combine for 64.6% of all 3,721 records in scope, and the top ten reach 73.0%. New entrants are filing into a field where core architecture claims are already occupied by a handful of large processor and memory vendors.
Based on the full assignee ranking
Momentum
-32% (2021→2024)
leading assignee filings
Growth has cooled from its 2022 peak, not stalled
Filings hit 633 in 2022 before easing to 396 by 2024, a 32% drop from the 2021 level. Several of the largest assignees show sharp year-over-year declines in the latest tracked year, though publication lag makes the newest years an unreliable read on direction.
Filing-year trend, 2017-2026
Composition
81.8% carry G06F
of 3,721 records
Architecture claims dominate over memory-cell claims
G06F class coverage at 81.8% signals that most filings are architectural or system-level rather than device-physics claims on the memory cell itself; G11C, the classic memory-device class, appears on only 10.2% of records.
IPC subclass shares, record-level
Citations
510 citations
top-cited record
Foundational smart-memory claims still anchor the field
The most-cited record in scope, on using smart memories in computing, carries 510 citations — well ahead of the next tier. Later system-level filings citing it show the field building outward from a small set of foundational architecture patents rather than replacing them.
Citation counts within this corpus
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 in-memory computing patent landscape, with the prior art for and against each one.
Source: Patsnap Eureka. Co-assignee relationships and derived observations. Derived from a Patsnap search on In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Players
Who holds the claim space, and what remains open
The ranked leaders span storage, semiconductor and systems vendors; a long tail of single- or few-filing entrants sits below them.
Leader
1,510 records
leading assignee
One filer sits well ahead of the field
The top-ranked assignee holds 1,510 records, far outdistancing fifth place at 116 and tenth place at 51 — the drop-off from first to fifth is steep even within the concentrated top group.
Assignee ranking, family count
Momentum shift
-84% to -100% YoY
among several leaders
Several top filers pulled back sharply in the latest year
Multiple leading assignees show large year-over-year declines in the most recent tracked year, some approaching zero new filings. Given publication lag, this looks more like a reporting gap than a genuine exit from the field.
Latest-year momentum by assignee
Collaboration
10 co-assignee pairs
identified
Co-filing is limited and concentrated in a few pairs
Only ten co-assignee pairs appear across the dataset, with the strongest pairing linking a major Chinese technology vendor and a leading Chinese university. Most filers in this field patent independently rather than through joint ventures.
Co-assignee pair analysis
Recent-year filing momentum by assignee
| Assignee | Recent year | YoY |
|---|---|---|
| Pure Storage, Inc. | 21 | -84% |
| McAfee LLC | 1 | -80% |
| Samsung Electronics Co., Ltd. | 1 | -94% |
| QUADRIC IO INC | 1 | -67% |
| Intel Corporation | 0 | -100% |
| International Business Machines Corporation | 0 | — |
| Ultrata LLC | 0 | — |
| Micron Technology, Inc. | 0 | -100% |
Source: Patsnap Eureka. Assignee-level momentum. Derived from a Patsnap search on In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
What's next
Where to take this analysis
The dataset points to a field with an occupied core and thinner edges — the next steps depend on where a given team sits relative to that line.
Map a specific claim against the leaders' portfolios
With one assignee holding 1,510 records, a targeted novelty check against that portfolio before drafting saves rework later.
Run a claim comparison in EurekaTrack the under-claimed IPC branches
H03K, H03M and H01L each sit under 2% of records — worth monitoring for filing activity as the core architecture classes fill up.
Set up class-level monitoring in EurekaWatch for the 2025-2026 correction
Because publication lag understates the newest years, re-check the trend once 2025 filings fully post before concluding the field is shrinking.
Build a filing-trend alert in EurekaSource: Patsnap Eureka. Forward-looking reading of the same dataset. Derived from a Patsnap search on In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
FAQ
Frequently asked questions
One assignee leads the ranked field with 1,510 records, well ahead of the rest of the top group — fifth place holds 116 and tenth place holds 51. The top five assignees combined account for 64.6% of all 3,721 records in scope, and the top ten reach 73.0%. That means most of the architectural claim space sits with a small number of storage, semiconductor and systems vendors, and any new filer should expect to design around at least one of them.
Filings grew strongly from 76 in 2017 to a peak of 633 in 2022, then eased to 396 by 2024 — a 32% decline over the 2021-2024 window. That is the most recent period that can be read reliably, because publication lags filing by roughly 18 months and 2025-2026 figures are still incomplete. Read the recent slowdown as a cooling from an unusually high 2022 peak rather than a definitive end to growth.
G06F (electric digital data processing) covers 81.8% of the 3,721 records in scope, making it the dominant class by a wide margin. H04L (digital transmission, 18.5%), G06N (AI-model computing, 11.4%) and G11C (static and digital memories, 10.2%) follow. Because a single record can carry several IPC classes, these percentages overlap rather than sum to 100%, and each should be read against the full record total, not against one another.
The classes with the lowest record shares — H03K (pulse technique and logic circuits), H03M (coding and code conversion), H04W (wireless networks) and H01L (semiconductor devices), each under 2% of records — mark branches that are comparatively thin relative to the core G06F architecture claims. That does not guarantee an easy filing, but it does mean the claim density there is lower, which is where a freedom-to-operate search is more likely to surface open space rather than blocking prior art.
The most-cited record in this corpus is US6807614B2, covering the use of smart memories in computing, with 510 citations, followed by US7546438B2 on algorithm mapping and specialized instructions for smart memory computing at 432 citations. High citation counts inside a searched corpus tend to favour older filings that have simply had more time to accumulate references, so these figures signal foundational influence rather than current commercial weight.
Answers are grounded in the same dataset. Derived from a Patsnap search on In-Memory Computing Patent Landscape covering 2015–2026, data cut-off 2026-08-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Research In-Memory Computing Patent Landscape in depth with Eureka
Go past this page: query the whole in-memory 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.