Artificial Intelligence Computing Patent Landscape 2026
Artificial Intelligence Computing Patent Landscape in 2026
The AI computing patent field, encompassing 18,413 patent families, is past its 2020 peak with annual volume easing on a multi-year basis, though the space remains highly active and moderately concentrated among a small group of large technology incumbents. IBM leads with the largest portfolio, while Baidu and Qualcomm show the strongest recent filing momentum among the top tier.
IBM leads a moderately concentrated field with strong challenger pressure
IBM holds the top position with 964 patent families, followed closely by Google (766) and Microsoft (745), forming a tight leading trio. Robert Bosch (639) and Baidu (467) round out the top five, demonstrating that the competitive landscape spans both pure-software technology giants and industrial conglomerates.
The top five filers account for 31% of the combined output of the hundred largest filers — a meaningful but not dominant concentration. This indicates that while incumbents have scale advantages, a broad second tier of challengers from Qualcomm and NVIDIA to Accenture and Capital One remains actively competitive.
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 1 | International Business Machines Corporation | 964 | |
| 2 | Google LLC | 766 | |
| 3 | Microsoft Technology Licensing LLC | 745 | |
| 4 | Robert Bosch GmbH | 639 | |
| 5 | BEIJING BAIDU NETCOM SCI & TECH CO LTD | 467 | |
| 6 | Strong Force IoT Portfolio 2016 LLC | 388 | |
| 7 | Qualcomm Incorporated | 383 | |
| 8 | NVIDIA Corporation | 376 | |
| 9 | Accenture Global Solutions Ltd. | 253 | |
| 10 | Strong Force TX Portfolio 2018 LLC | 238 |
| # | Applicant | Patent families | Share |
|---|---|---|---|
| 11 | Capital One Services LLC | 233 | |
| 12 | Huawei Technologies Co., Ltd. | 201 | |
| 13 | D5AI LLC | 193 | |
| 14 | Amazon Technologies Inc. | 184 | |
| 15 | Intel Corporation | 183 | |
| 16 | Bank of America Corporation | 178 | |
| 17 | Samsung Electronics Co., Ltd. | 149 | |
| 18 | Salesforce Inc. | 147 | |
| 19 | Oracle International Corporation | 143 | |
| 20 | Adobe Inc. | 138 |
IBM’s depth across multiple AI model sub-classes, combined with its active co-filing relationships with university partners, suggests it is building a portfolio designed for both licensing leverage and technical breadth. The presence of portfolio assertion entities (Strong Force family) in the top ten signals that IP monetization is already a feature of this landscape.
The most recent 18–24 months of filing data are subject to publication lag and will undercount true activity; apparent declines in 2024–2026 figures should not be treated as definitive trend signals. Longer-window growth, applicant concentration, and technology-route coverage are therefore more reliable signals than the latest-year bar alone.
Filing volume eased from a 2020 peak; AI model computing dominates the technology mix
The annual filing trend reveals a field that grew rapidly from 2017 to a 2020 peak and has since moderated, while the technology composition chart shows an overwhelming concentration in core AI model computing classes with a long tail of application-domain branches.
Annual filing trend
Filings rose steeply from 875 in 2017 to a peak of 3,146 in 2020, then declined through 2022 and beyond. Years 2024–2026 are materially under-counted due to publication lag and should not be read as a continuation of the decline.
↗ Hover for values · click a bar to ask EurekaTechnology composition
G06N (Computing based on AI models) is the overwhelmingly dominant class, reflecting the core algorithmic focus of the corpus. G06F (Electric digital data processing) and G06Q (Business and commerce data processing) form a secondary tier, while healthcare informatics (G16H), control systems (G05B), and medical diagnosis (A61B) illustrate significant application-domain extension.
↗ 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.
Artificial intelligence computing device and relat…
The invention provides an artificial intelligence computing device and a related product. The artificial intelligence computing device is used for executing machine learning computation. According to the device of the invention, for the instructions in the more than two instruction sets forming the loop body, the same operation code in the operation code… (excerpt from the patent abstract)
| # | Patent | Citations |
|---|---|---|
| 1 | Methods and systems for data collection, learning,… | 960 |
| 2 | System and Method for Synthetic Interaction with U… | 898 |
| 3 | Semisupervised autoencoder for sentiment analysis | 787 |
| 4 | Distributed Machine Learning Systems, Apparatus, a… | 755 |
| 5 | Platform for facilitating development of intellige… | 721 |
| 6 | Intelligent vibration digital twin systems and met… | 694 |
| 7 | Systems and methods for crowdsourcing information … | 622 |
| 8 | Mental Model Elicitation Device (MMED) Methods and… | 590 |
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. Some patent titles may be shown in their original, non-English language where an accurate translation could not be guaranteed.
What the AI computing patent structure means for R&D investment decisions
Four structural dimensions — maturity, competitive concentration, collaboration networks, and geographic reach — each carry distinct implications for where new entrants and incumbents should focus resources.
Post-peak but still substantive: a maturing growth field
The lifecycle stage is classified as Decline, with annual filings easing back from the 2020 peak of 3,146. The field remains substantive in absolute volume, and the multi-year corpus of 18,413 patent families signals deep prior-art density. New entrants should expect significant freedom-to-operate constraints in core G06N classes and should target differentiated sub-domains rather than broad AI model methods.
Post-peak · 2020 peakModerate top-tier concentration with a broad competitive second tier
The top five filers hold 31% of the hundred largest filers’ combined total, leaving 69% distributed across a wide challenger field. Portfolio assertion entities (Strong Force IoT Portfolio 2016, Strong Force TX Portfolio 2018) occupying top-ten positions signal active licensing pressure. R&D teams should map their proposed claim space against these portfolios before investing heavily in adjacent application domains.
31% top-5 shareIBM dominates co-filing networks; Bosch–Carnegie Mellon leads external partnerships
The most active co-filing pair is Robert Bosch and Carnegie Mellon University with 28 joint families, pointing to a strong industry-academia bridge in applied AI for physical systems. IBM co-files extensively with its own international subsidiaries (IBM China, IBM UK, IBM Germany) and with MIT (10 families) and Rensselaer Polytechnic Institute (4 families). Teams seeking academic partnership signals should note that Carnegie Mellon and MIT are the most active university nodes in this network.
Bosch–CMU · IBM–MITUS-centric filing with moderate EPO and PCT international reach
The United States is the dominant filing jurisdiction by a wide margin, reflecting both the location of leading applicants and the strategic importance of the US market. Europe (EPO) and WIPO (PCT) form a meaningful secondary tier, while India and Canada represent a modest but growing next layer. China’s relatively low count in this corpus likely reflects that Chinese-origin applicants (e.g., Baidu) file primarily through domestic channels not fully captured here, and should be assessed through a separate CNIPA search.
US-dominant · EPO secondaryGo 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 |
|---|---|---|
| Robert Bosch GmbH | Carnegie Mellon University | 28 |
| International Business Machines Corporation | IBM China Co., Ltd. | 22 |
| International Business Machines Corporation | IBM UNITED KINGDOM LTD INTELLECTUAL PROPERTY DEPAR… | 12 |
| International Business Machines Corporation | Massachusetts Institute of Technology | 10 |
| International Business Machines Corporation | IBM DEUTSCHLAND GMBH | 7 |
| International Business Machines Corporation | Rensselaer Polytechnic Institute | 4 |
| International Business Machines Corporation | Board of Trustees of the University of Illinois | 3 |
| Robert Bosch GmbH | Cariiad Ltd. | 3 |
| International Business Machines Corporation | University of Massachusetts | 2 |
| International Business Machines Corporation | IHI Corporation | 2 |
Co-filing pairs, ranked by the number of jointly-filed patent families.
IBM anchors the field; Baidu and Qualcomm show the sharpest recent momentum
The top applicants split broadly into two groups: established incumbents with large but moderating portfolios, and challengers with accelerating recent activity. Technology focus across all leaders centers on G06N AI model computing, with differentiation in application layers.
International Business Machines Corporation
IBM leads with 964 patent families, concentrated in G06N20 (machine learning methods, 695 families) and G06N3 (neural computing, 379 families). Recent filing momentum has contracted sharply (trend: -78% vs prior period), suggesting the portfolio is being selectively maintained rather than aggressively expanded. IBM’s co-filing network with MIT, Rensselaer Polytechnic, the University of Illinois, and the University of Massachusetts reinforces its position as a central node in the broader AI computing ecosystem.
964 patent familiesBeijing Baidu Netcom Science & Technology Co., Ltd.
Baidu ranks fifth overall with 467 patent families and is the fastest-growing major filer in the corpus, with recent filings up 119% versus the prior period — the strongest positive trend among tracked leaders. Its technology focus spans G06N3 (neural computing, 306 families), G06N20 (165 families), and G06F40 (natural language processing, 78 families), indicating a concentrated push in deep learning and NLP-adjacent methods. Qualcomm (383 families, +133% recent trend) presents a comparable acceleration story, particularly in on-device neural inference, and merits equal attention as a fast-moving challenger.
467 patent families| Applicant | Recent (3 yrs) | Trend |
|---|---|---|
| International Business Machines Corporation | 131 | ▼ -78% |
| Google LLC | 165 | ▬ -1% |
| Robert Bosch GmbH | 187 | ▼ -33% |
| Microsoft Technology Licensing LLC | 146 | ▼ -22% |
| Beijing Baidu Netcom Science & Technology Co., Ltd. | 280 | ▲ +119% |
| Strong Force IoT Portfolio 2016 LLC | 53 | ▼ -68% |
| NVIDIA Corporation | 108 | ▼ -41% |
| Qualcomm Incorporated | 177 | ▲ +133% |
Under-served branches adjacent to the AI computing core
Several IPC classes appear at relatively lower share within the corpus despite having clear technical linkage to AI computing methods; these represent areas where prior-art density is lower and differentiated filing may encounter less crowding.
G16B · Bioinformatics & computational biology
G16B accounts for a relatively small share of the corpus despite the well-established application of AI models to genomic sequence analysis, protein structure prediction, and drug-target interaction modeling. The technical pathway is direct — G06N3 neural architectures are already applied in this domain — and the lower filing density suggests that integrated AI-bioinformatics claims combining algorithmic novelty with biological data structures may encounter less prior-art crowding than equivalent claims in the core G06N class.
Search this in Eureka →G16Y · IoT data processing
G16Y (IoT data processing) carries a notably low count relative to the corpus size, even though AI-driven edge inference and federated learning over IoT sensor networks are active engineering areas. Given that Strong Force IoT Portfolio 2016 is already a top-ten filer with a G05B/G06N focus, this adjacent branch warrants a careful freedom-to-operate review before entry, but the relatively sparse G16Y-specific filing suggests there may be space for claims tightly scoped to AI inference protocols on constrained IoT endpoints.
Search this in Eureka →How leading applicants differ across technology routes
Strength of each leader across the main technology routes.
| Player | G06N 20 · Computing based on AI models | G06N 3 · Computing based on AI models | G06N 5 · Computing based on AI models | G06K 9 · Data recognition & presentation | G06F 16 · Electric digital data processing |
|---|---|---|---|---|---|
| International Business Machines Corporation | Strong · 695 | Strong · 379 | Moderate · 226 | Moderate · 181 | Moderate · 180 |
| Strong Force IoT Portfolio 2016 LLC | Strong · 370 | Strong · 407 | Strong · 379 | Strong · 246 | Absent |
| Google LLC | Strong · 336 | Strong · 396 | Emerging · 59 | Emerging · 61 | Emerging · 71 |
| Strong Force TX Portfolio 2018 LLC | Strong · 219 | Strong · 183 | Strong · 127 | Moderate · 84 | Strong · 161 |
| Microsoft Technology Licensing LLC | Strong · 349 | Strong · 223 | Emerging · 67 | Emerging · 54 | Moderate · 72 |
| Robert Bosch GmbH | Strong · 214 | Strong · 418 | Emerging · 45 | Emerging · 78 | Absent |
| Beijing Baidu Netcom Science & Technology Co., Ltd. | Strong · 165 | Strong · 306 | Absent | Absent | Emerging · 59 |
Frequently asked questions
The corpus covers 18,413 patent families. This total is the base from which applicant rankings, lifecycle assessment, and technology composition analysis are derived.
International Business Machines Corporation (IBM) leads with 964 patent families, ahead of Google (766) and Microsoft (745). Together these three form the leading tier.
The field is classified as post-peak. Annual filings reached their highest point in 2020 and have eased since. The most recent years (2024–2026) are subject to publication lag and undercount true activity, so the full extent of any moderation will only be clear once those filings publish.
The United States is by far the dominant filing jurisdiction. Europe (EPO) and WIPO (PCT) form a secondary tier, with India and Canada as a next layer. Note that these are patent-record level counts and a family filed in multiple jurisdictions will appear under each.
Among the tracked top applicants, Qualcomm shows the strongest recent upward trend (+133% versus the prior period), followed closely by Baidu (+119%). Google is roughly flat (-1%), while IBM (-78%), Strong Force IoT Portfolio 2016 (-68%), and NVIDIA (-41%) have contracted significantly in recent filings.
The most active co-filing pair is Robert Bosch and Carnegie Mellon University with 28 jointly filed families. IBM co-files with MIT (10 families), Rensselaer Polytechnic Institute (4 families), the University of Illinois (3 families), and the University of Massachusetts (2 families), making IBM the central hub of the university collaboration network in this corpus.
Map the patent landscape for your own AI computing focus area
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.