Federated & Privacy-Preserving ML Patent Landscape
Federated & Privacy-Preserving ML Patent Landscape in 2026
The federated and privacy-preserving ML field has reached 19,996 patent families and is in an active growth phase, with annual volume still rising and the multi-year window up 123%. Huawei leads a moderately concentrated field, but a broad tier of telecom, hyperscaler, and fintech challengers keeps the competitive structure genuinely open.
Huawei leads a moderately concentrated field with strong challengers close behind
Huawei Technology leads all applicants with 586 patent families, followed closely by Qualcomm (490), Ericsson (487), and Google (485) — a remarkably compressed top-four in which fewer than 100 families separate first from fourth place.
The top five filers collectively account for 24% of the hundred largest filers’ combined total, indicating moderate rather than dominant concentration. No single entity commands an insurmountable lead, and a second tier — IBM, WeBank, Alipay, NVIDIA, Beijing University of Posts and Telecommunications — keeps competitive pressure broad.
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 1 | Huawei Technologies Co., Ltd. | 586 | |
| 2 | Qualcomm Incorporated | 490 | |
| 3 | Telefonaktiebolaget LM Ericsson | 487 | |
| 4 | Google LLC | 485 | |
| 5 | Samsung Electronics Co., Ltd. | 331 | |
| 6 | International Business Machines Corporation | 303 | |
| 7 | WeBank Co., Ltd. | 285 | |
| 8 | Alipay (Hangzhou) Information Technology Co., Ltd. | 260 | |
| 9 | NVIDIA Corporation | 256 | |
| 10 | BEIJING UNIV OF POSTS & TELECOMM | 234 |
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 11 | Tencent Technology (Shenzhen) Co., Ltd. | 226 | |
| 12 | Nokia Technologies Oy | 195 | |
| 13 | NANJING UNIV OF POSTS & TELECOMM | 166 | |
| 14 | Xidian University | 165 | |
| 15 | China Mobile Communications Group Co., Ltd. | 148 | |
| 16 | Chongqing University OF POSTS & TELECOMM | 139 | |
| 17 | LG Electronics Inc. | 139 | |
| 18 | InterDigital Patent Holdings, Inc. | 135 | |
| 19 | Zhejiang University | 134 | |
| 20 | Ping An Technology (Shenzhen) Co., Ltd. | 133 |
The near-parity among the top four suggests that leadership could shift as investment cycles evolve. The presence of both telecom infrastructure players (Ericsson, Qualcomm) and hyperscalers (Google) alongside hardware (NVIDIA) and fintech specialists (WeBank, Alipay) signals that no single application vertical has yet claimed the field.
Data for the most recent 18–24 months is subject to publication lag and likely understates current activity; recent-period counts should be read as minimums. Longer-window growth, applicant concentration, and technology-route coverage are therefore more reliable signals than the latest-year bar alone.
Filings are still rising and AI-model computing dominates the technology mix
The annual filing trend and technology composition together reveal a field in active growth whose core IP is concentrated in AI and data-processing classes, with meaningful secondary activity in communications and commercial applications.
Annual filing trend
Filings grew steeply from 70 in 2017 to over 3,000 annually by 2022 and continued climbing through 2024. The 2025–2026 figures appear elevated but are partially inflated by pending publications; treat them as indicative minimums rather than final counts. The multi-year window reflects 123% growth, consistent with the field’s active growth lifecycle stage.
↗ Hover for values · click a bar to ask EurekaTechnology composition
G06N (computing based on AI models) is the commanding primary class. G06F (electric digital data processing) and H04L (digital information transmission) form a substantial secondary layer, reflecting the distributed-systems and cryptographic-protocol engineering that underpins federated architectures. G06Q (business and commerce) and H04W (wireless networks) indicate meaningful application-layer activity in fintech and mobile edge computing respectively.
↗ 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.
Resource-limited federated learning using dynamic …
A computer-implemented method, a computer program product, and a computer system for resource-limited federated learning using dynamic masking. A server in federated machine learning evaluates resources of respective agents in the federated machine learning to determine capacities of model training by the respective agents. The server masks weights of a… (excerpt from the patent abstract)
| # | Patent | Citations |
|---|---|---|
| 1 | Technologies for distributing iterative computatio… | 363 |
| 2 | Technologies for distributing gradient descent com… | 332 |
| 3 | Multi-view deep neural network for lidar perception | 254 |
| 4 | Vehicle-data analytics | 252 |
| 5 | Application Development Platform and Software Deve… | 243 |
| 6 | 一种联邦学习训练数据隐私性增强方法及系统 | 206 |
| 7 | Data Reproducibility Using Blockchains | 206 |
| 8 | Privacy-preserving machine learning | 188 |
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 active growth, moderate concentration, multi-sector collaboration, and. China-heavy geography creates distinct strategic considerations for any team entering or expanding in this space.
Active growth — annual volume still rising from a low 2017 base
The lifecycle stage is Growth, with annual filings still climbing and a multi-year growth rate of 123%. The field moved from 70 families in 2017 to over 3,000 per year by 2022 and has not yet peaked. This means core architectural approaches are still being staked out and first-mover IP positions remain achievable in less-covered sub-domains.
Growth stageModerate concentration with a compressed top four and a deep second tier
The top five filers hold 24% of the hundred largest filers’ combined total — moderate, not dominant. The gap between first-place Huawei (586 patent families) and fourth-place Google (485 patent families) is narrow enough that portfolio momentum, not raw stock, will determine leadership in the next cycle. A deep second tier of eight-plus active players limits any single entity’s ability to wall off the field.
Competitive openIBM is the most active co-filer; university-industry pairs are common in China
IBM (international entity) co-files most actively with IBM China (13 joint families) and Rensselaer Polytechnic Institute (13 joint families), and also collaborates with IBM UK and IBM Germany. In China, Beijing University of Posts and Telecommunications co-files with State Grid Corporation of China (7 families) and State Grid Zhejiang (6 families). Huawei co-files with Tsinghua University (5 families) and Nanjing University (4 families), reflecting a broader university-industry ecosystem underpinning Chinese applicants.
Ecosystem activeChina dominates filings; US and India are the main secondary offices
China accounts for the largest share of patent records, followed by the United States and India — a distribution that reflects both domestic regulatory incentives and the heavy presence of Chinese technology companies and universities in the applicant ranking. WIPO PCT and EPO filings indicate meaningful global protection strategies among top applicants. Teams seeking freedom-to-operate should prioritize clearance in China, the US, and India as a minimum baseline.
China-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 |
|---|---|---|
| International Business Machines Corporation | IBM China Co., Ltd. | 13 |
| International Business Machines Corporation | Rensselaer Polytechnic Institute | 13 |
| International Business Machines Corporation | IBM UNITED KINGDOM LTD INTELLECTUAL PROPERTY DEPAR… | 11 |
| Beijing University of Posts and Telecommunications | State Grid Corporation of China | 7 |
| Samsung Electronics Co., Ltd. | Korea Advanced Institute of Science and Technology (KAIST) | 6 |
| Beijing University of Posts and Telecommunications | State Grid Zhejiang Electric Power Research Institute | 6 |
| Huawei Technologies Co., Ltd. | Tsinghua University | 5 |
| International Business Machines Corporation | IBM DEUTSCHLAND GMBH | 5 |
| Beijing University of Posts and Telecommunications | State Grid Liaoning Electric Power Co., Ltd. | 5 |
| Huawei Technologies Co., Ltd. | Nanjing University | 4 |
Co-filing pairs, ranked by the number of jointly-filed patent families.
Huawei and Qualcomm lead with divergent momentum trajectories
The top applicants share a common AI-model computing core but diverge sharply in secondary focus areas and recent filing momentum, creating distinct competitive profiles.
Huawei Technology
Huawei leads the field with 586 patent families and a recent-period trend of +79%, indicating sustained and accelerating investment. Its portfolio is centered on AI-model computing (G06N 20 and G06N 3) with a secondary layer in network management (H04L 41), consistent with its integrated hardware-software-infrastructure positioning. The +79% momentum suggests Huawei is actively extending its lead rather than resting on an existing stock.
families: 586Qualcomm
Qualcomm ranks second with 490 patent families and shows the strongest momentum among the top four at +108%, making it the fastest-growing major filer in the field. Its primary emphasis is neural-network computing (G06N 3 and G06N 20) with a secondary focus on wireless networks (H04W 24), reflecting an edge-inference and on-device federated learning strategy. At this growth rate, Qualcomm is positioned to challenge Huawei’s lead over the next filing cycle.
families: 490| Applicant | Recent (3 yrs) | Trend |
|---|---|---|
| Huawei Technologies Co., Ltd. | 346 | ▲ +79% |
| Telefonaktiebolaget LM Ericsson | 212 | ▼ -15% |
| Google LLC | 205 | ▲ +42% |
| Qualcomm Incorporated | 295 | ▲ +108% |
| Samsung Electronics Co., Ltd. | 230 | ▲ 3.2× vs prior 3-yr |
| International Business Machines Corporation | 105 | ▼ -45% |
| WeBank Co., Ltd. | 31 | ▼ -88% |
| Alipay (Hangzhou) Information Technology Co., Ltd. | 103 | ▼ -31% |
Under-served adjacent branches worth monitoring
Several IPC classes with plausible technical relevance to federated and privacy-preserving ML carry lower relative shares, suggesting they may be under-exploited relative to their application potential.
G16H · Healthcare Informatics
Healthcare informatics holds a 3% share among the top branches, despite federated learning being the canonical solution to multi-hospital data silos under HIPAA and GDPR constraints. The gap between the evident technical fit — training diagnostic models without centralizing patient data — and the relatively sparse patent activity suggests that healthcare-specific federated protocols (e.g., differential-privacy guarantees for clinical NLP, federated survival analysis) remain under-staked. Entry paths include filing around vertical-specific aggregation protocols, consent-preserving data-sharing architectures, and federated model auditing for regulated environments.
Search this in Eureka →G16B · Bioinformatics
Bioinformatics (G16B) appears with only 47 records in the corpus — one of the sparsest branches relative to its technical potential. Genomic and multi-omics data represent a high-sensitivity, legally restricted category where federated and secure multi-party computation approaches are directly applicable, yet IP activity is minimal. The combination of sparse filing, strong regulatory pull (genomic privacy regulations expanding globally), and clear technical demand makes this an adjacent branch worth watching for early-mover positioning.
Search this in Eureka →How leading applicants differ by technology route
Strength of each leader across the main technology routes.
| Player | G06N 20 · Computing based on AI models | G06N 3 · Computing based on AI models | G06F 21 · Electric digital data processing | H04L 9 · Digital information transmission | G06F 18 · Electric digital data processing |
|---|---|---|---|---|---|
| Huawei Technologies Co., Ltd. | Strong · 329 | Strong · 312 | Absent | Emerging · 35 | Absent |
| Telefonaktiebolaget LM Ericsson | Strong · 269 | Strong · 303 | Absent | Moderate · 75 | Absent |
| Google LLC | Strong · 281 | Strong · 234 | Emerging · 50 | Emerging · 51 | Absent |
| Qualcomm Incorporated | Strong · 199 | Strong · 339 | Absent | Absent | Absent |
| International Business Machines Corporation | Strong · 183 | Strong · 108 | Moderate · 52 | Moderate · 78 | Absent |
| Samsung Electronics Co., Ltd. | Strong · 208 | Strong · 155 | Absent | Absent | Absent |
| Beijing University of Posts and Telecommunications | Strong · 131 | Strong · 114 | Moderate · 56 | Absent | Moderate · 56 |
Frequently asked questions
The corpus covers 19,996 patent families in scope globally. Annual filings grew from 70 in 2017 to over 3,000 per year by 2022 and have continued to climb, with a multi-year growth rate of 123%. The field is in an active growth lifecycle stage.
Huawei Technology leads with 586 patent families and a recent-period filing trend of +79%, indicating sustained acceleration. Qualcomm (490 patent families, +108% trend) is the fastest-growing major filer among the top four.
The top five filers account for 24% of the hundred largest filers’ combined total, indicating moderate concentration. The gap between first-place Huawei and fourth-place Google is fewer than 100 patent families, so the leadership position is not entrenched.
China records the highest volume of patent records, followed by the United States and India. WIPO PCT and EPO filings indicate that leading applicants are pursuing broad international protection. Teams assessing freedom-to-operate should prioritize China, the US, and India as a minimum.
G06N (computing based on AI models) is the dominant class, reflecting the core algorithmic nature of federated learning. G06F (electric digital data processing) and H04L (digital information transmission) form a substantial secondary layer covering data engineering and cryptographic-protocol work.
Healthcare informatics (G16H) and bioinformatics (G16B) are among the branches with lower relative filing density despite strong technical and regulatory rationale for federated and privacy-preserving approaches. These are adjacent branches worth monitoring, though they should not be treated as validated commercial opportunities without further competitive analysis.
Ready to map your own federated ML 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.