Edge AI Accelerators Patent Landscape 2026
Edge AI Accelerators Patent Landscape in 2026
Samsung Electronics commands the largest position among the top hundred filers and, together with Qualcomm, accounts for a dominant share of the concentrated leader tier. The field is in active growth — annual filings have risen consistently since 2017, with the most recent years still accumulating under publication lag.
Samsung and Qualcomm anchor a steeply concentrated leader tier
Samsung Electronics ranks first with 2,689 patent families, more than 1.5 times the second-ranked Qualcomm (1,773 patent families). Together they represent by far the largest individual positions in the corpus of 11,282 patent families in scope.
The top five filers — Samsung, Qualcomm, Huawei, Intel, and DeepX — account for 63% of the combined output of the hundred largest filers. This degree of concentration is characteristic of a field where foundational hardware and algorithm IP has consolidated early around a small set of semiconductor and platform companies.
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 1 | Samsung Electronics Co., Ltd. | 2,689 | |
| 2 | Qualcomm Incorporated | 1,773 | |
| 3 | Huawei Technologies Co., Ltd. | 416 | |
| 4 | Intel Corporation | 314 | |
| 5 | DeepX Co., Ltd. | 275 | |
| 6 | ARM Ltd. | 195 | |
| 7 | VIA Alliance Semiconductor Co., Ltd. | 160 | |
| 8 | D5AI LLC | 153 | |
| 9 | Digital Global Systems Inc. | 140 | |
| 10 | Microsoft Technology Licensing LLC | 132 |
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 11 | SEOUL NATIONAL UNIVERSITY R&DB FOUNDATION | 108 | |
| 12 | SAS Institute Inc. | 93 | |
| 13 | Siemens AG | 83 | |
| 14 | Google LLC | 80 | |
| 15 | Huawei Cloud Computing Technologies Co., Ltd. | 78 | |
| 16 | Qualcomm Technologies Inc. | 76 | |
| 17 | Micron Technology Inc. | 72 | |
| 18 | NVIDIA Corporation | 71 | |
| 19 | Rebellions Inc. | 57 | |
| 20 | Meta Platforms Technologies LLC | 56 |
Samsung’s scale reflects both its broad semiconductor portfolio and an active co-filing program with Korean universities, while Qualcomm’s 122% momentum surge signals a recent strategic acceleration. The tier gap between these two leaders and the rest of the top ten is large enough that new entrants face a significant catch-up challenge in core neural-network compute IP.
Filings from approximately the last 18–24 months are under-counted due to patent-publication lag; the apparent 2026 figure in particular understates true activity. Longer-window growth, applicant concentration, and technology-route coverage are therefore more reliable signals than the latest-year bar alone.
Consistent annual growth with AI-model compute as the dominant technology class
The filing trend and technology composition charts together show a field still in expansion, anchored by neural-network compute IP but extending into adjacent application domains at the periphery.
Annual filing trend
Annual filings grew from 112 in 2017 to a recorded high of 1,994 in 2025, reflecting a 42% increase over the recent window. The 2026 bar is materially incomplete due to publication lag and should not be read as a slowdown; the growth trajectory established since 2017 remains intact.
↗ Hover for values · click a bar to ask EurekaTechnology composition
G06N (AI-model computing) is overwhelmingly dominant. G06F (digital data processing) is the second-largest branch, followed at some distance by G06T (image processing), G06V (image/video recognition), and H04L (digital transmission) — reflecting the deployment of accelerators in vision and connectivity pipelines. Lower-volume branches such as G11C (memory), H03K (logic circuits), and H01L (semiconductor devices) mark hardware-layer activity that remains relatively sparse compared to the algorithm-layer core.
↗ Hover for values · click a bar to ask EurekaHighly cited patent families surfaced by the query
Citation-heavy patent families returned by the query. Use this section as citation context, not as a curated list of the most topic-specific patents.
Multiomoal emotion recognition system with edge ai…
The present invention relates a multimodal emotion recognition system with edge AI accelerator. The system employs a Fuzzy Long-Short Recurrent Convolutional Network (Fuzzy LRCN) for multiple emotion classifications. The algorithm strategically utilizes only eight channels of electroencephalogram (EEG) signals and combines electrocardiogram (ECG) and… (excerpt from the patent abstract)
| # | Patent | Citations |
|---|---|---|
| 1 | Technologies for distributing iterative computatio… | 363 |
| 2 | Technologies for distributing gradient descent com… | 332 |
| 3 | System, method, and apparatus for providing dynami… | 298 |
| 4 | Fixed point neural network based on floating point… | 260 |
| 5 | Adaptive selection of artificial neural networks | 251 |
| 6 | Bit width selection for fixed point neural networks | 249 |
| 7 | Application Development Platform and Software Deve… | 243 |
| 8 | Method and apparatus for automated design of compl… | 241 |
Ranked by total forward citations. Citation counts favour older and broadly cited patent families, and broad or adjacent patents may appear when they match the search scope. Treat this section as citation context, not as a curated list of the most topic-specific patents.
What the competitive structure means for R&D strategy
The combination of high concentration, a growth life-cycle stage, active industry–university collaboration, and US-led jurisdiction coverage defines the strategic landscape for an entrant or challenger.
Growth stage with publication lag masking true pace
The field is classified as Growth: annual filings have risen consistently and the most recent years remain understated by publication lag. This means the window for establishing foundational positions is not closed, but the pace of incumbents’ filings is high enough that differentiated technical angles — rather than broad AI-compute coverage — are the realistic entry path for new filers.
Life-cycle: GrowthTwo-tier structure with a wide gap below the top pair
Samsung and Qualcomm together hold a position that no single challenger replicates. Huawei (416 patent families), Intel (314), and DeepX (275) form a mid-tier that is active but substantially smaller. Below rank five, portfolio sizes fall sharply, suggesting that targeted sub-domain strategies — rather than broad portfolio buildout — are more realistic for mid-tier and emerging players.
Top-5 share: 63%Samsung anchors a dense industry–university co-filing network
Samsung Electronics co-files most actively with the Seoul National University R&DB Foundation (95 joint patent families), followed by KAIST (35) and Sungkyunkwan University (18). ARM co-files with an entity translating as ‘Top Company’ (16 joint families). This pattern shows that academic partnerships are a meaningful source of novel IP in this space, particularly for hardware-algorithm co-design, and that Korean universities are disproportionately embedded in the ecosystem.
Co-filing network activeUS leads filings; India and EPO are the next-largest jurisdictions
The United States is the primary filing jurisdiction, followed by India and Europe (EPO). WIPO (PCT) filings indicate broad international protection strategies by leading applicants. China, Germany, Taiwan, and Japan form a secondary tier. The relatively high India count reflects both domestic filers and international applicants extending protection into a fast-growing deployment market. Coverage across Southeast Asia (Vietnam, Philippines, Singapore) is present but thin.
US-led, global reachGo beyond the landscape: Eureka’s TRIZ Solution agent breaks down an R&D problem and returns patented concept solutions, each with a technical approach and cited patent & literature evidence.
| Applicant | Collaborator | Co-filings |
|---|---|---|
| Samsung Electronics Co., Ltd. | Seoul National University R&DB Foundation | 95 |
| Samsung Electronics Co., Ltd. | Korea Advanced Institute of Science and Technology (KAIST) | 35 |
| Samsung Electronics Co., Ltd. | Research & Business Foundation Sungkyunkwan University | 18 |
| Samsung Electronics Co., Ltd. | Industry Academic Cooperation Foundation Yonsei University | 18 |
| ARM Ltd. | Top Company | 16 |
| Samsung Electronics Co., Ltd. | Nota Inc. | 12 |
| Samsung Electronics Co., Ltd. | POSTECH Academy-Industry Foundation | 10 |
| Samsung Electronics Co., Ltd. | Ulsan National Institute of Science and Technology (UNIST) | 8 |
| Samsung Electronics Co., Ltd. | University of Zurich | 8 |
| Samsung Electronics Co., Ltd. | 高丽大学校产学协力团 | 6 |
Co-filing pairs, ranked by the number of jointly-filed patent families.
Samsung and Qualcomm lead on volume; DeepX stands out on momentum
The top applicants differ in portfolio breadth, technology emphasis, and recent trajectory. Samsung leads on absolute volume; Qualcomm has accelerated sharply; DeepX shows the fastest growth rate from a specialist base.
Samsung Electronics
Samsung holds 2,689 patent families — the largest position in the field — with its portfolio concentrated in G06N 3 (neural-network computing), G06T 7 (image processing), and G06F 3 (digital data processing). Recent momentum is positive at +19%, and a co-filing program spanning at least seven Korean university partners reinforces its foundational coverage across both algorithm and hardware-adjacent layers.
families: 2,689Qualcomm
Qualcomm ranks second with 1,773 patent families and has recorded the strongest momentum among established leaders at +122% in the recent period, signalling a deliberate strategic acceleration. Its technical focus sits in G06N 3 (neural-network computing), G06N 20 (machine learning), and G06V 10 (image/video recognition), consistent with on-device inference use cases in mobile and automotive platforms.
families: 1,773| Applicant | Recent (3 yrs) | Trend |
|---|---|---|
| Samsung Electronics Co., Ltd. | 1,283 | ▲ +19% |
| Qualcomm Incorporated | 903 | ▲ +122% |
| Huawei Technologies Co., Ltd. | 156 | ▬ +1% |
| Intel Corporation | 118 | ▬ -3% |
| DeepX Co., Ltd. | 189 | ▲ 6.5× vs prior 3-yr |
| ARM Ltd. | 81 | ▼ -10% |
| D5AI LLC | 46 | ▼ -21% |
| Digital Global Systems Inc. | 71 | ▲ new entrant |
Under-served branches adjacent to the dominant compute core
Several IPC branches sit at the periphery of the core G06N cluster with relatively modest patent-record counts. These are observations of relative sparsity; where a branch also has clear technical relevance and a plausible entry path, it may represent an actionable adjacent area.
G06T · Image data processing & generation
With 1,439 patent records, image processing is the third-largest branch but holds only a 6% share of the technology composition — notably sparse given that vision inference is one of the primary edge-AI use cases. The technical gap between accelerator architecture and real-time image-pipeline optimization (super-resolution, low-light enhancement, hardware-aware compression) is well-defined, and the entry path via existing computer-vision IP frameworks is realistic for both semiconductor and system-level filers.
Search this in Eureka →H04L · Digital information transmission
H04L covers digital transmission protocols and network-layer processing and appears in 909 patent records — a 4% share — despite edge AI deployments being inherently communication-dependent (federated learning, model synchronization, telemetry). The intersection of accelerator design with communication-efficient inference and on-device model aggregation is technically active but remains under-filed relative to pure compute IP, offering a differentiated angle for filers operating at the edge–network boundary.
Search this in Eureka →How leading filers differ across technology sub-routes
Strength of each leader across the main technology routes.
| Player | G06N 3 · Computing based on AI models | G06N 20 · Computing based on AI models | G06F 9 · Electric digital data processing | G06V 10 · Image/video recognition | G06F 15 · Electric digital data processing |
|---|---|---|---|---|---|
| Samsung Electronics Co., Ltd. | Strong · 2,496 | Emerging · 238 | Emerging · 189 | Emerging · 206 | Emerging · 128 |
| Qualcomm Incorporated | Strong · 1,725 | Emerging · 190 | Emerging · 54 | Emerging · 123 | Emerging · 51 |
| Huawei Technologies Co., Ltd. | Strong · 321 | Absent | Emerging · 32 | Emerging · 26 | Emerging · 36 |
| Intel Corporation | Strong · 260 | Absent | Moderate · 69 | Emerging · 18 | Emerging · 41 |
| DeepX Co., Ltd. | Strong · 260 | Emerging · 17 | Emerging · 24 | Emerging · 18 | Emerging · 28 |
| Digital Global Systems Inc. | Strong · 140 | Strong · 140 | Absent | Absent | Absent |
| VIA Alliance Semiconductor Co., Ltd. (Shanghai Zhaoxin Integrated Circuit) | Strong · 158 | Absent | Moderate · 62 | Absent | Emerging · 20 |
Frequently asked questions
The analysis covers 11,282 patent families globally. The most recent 18–24 months are under-counted due to patent-publication lag, so the true current total is higher.
Samsung Electronics is the largest filer with 2,689 patent families, more than 1.5 times the second-ranked Qualcomm (1,773 patent families). The top five filers together account for 63% of the combined output of the hundred largest filers.
The United States is the leading jurisdiction, followed by India and Europe (EPO). WIPO (PCT) filings are the fourth-largest route, indicating that leading applicants are pursuing broad international protection strategies.
The field is assessed as Growth stage. Annual filings have risen from 112 in 2017 to 1,994 in 2025, and the 42% recent-window growth rate confirms continued expansion. Publication lag means the most recent year figures understate actual activity.
Qualcomm shows the strongest momentum among established leaders at +122% in the recent period. DeepX (a specialist AI chip company) stands out with a 6.5× increase versus its prior three-year baseline, though from a smaller absolute base. Digital Global Systems is a new entrant with no meaningful prior-period count.
G06T (image data processing) holds a 6% share and G06V (image/video recognition) a 5% share despite edge-AI vision inference being a primary deployment use case. H04L (digital information transmission) at 4% is sparse relative to the communication dependency of edge deployments. These branches represent areas of relative sparsity within the broader field.
Ready to map your own Edge AI Accelerators landscape?
Join 18,000+ innovators using PatSnap Eureka to map any technology landscape: search 2B+ patents and papers, surface key assignees, and generate a report like this in minutes.
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.