Non-Terrestrial Network AI Patents: Who Leads, Trends 2026
- Filing peaked in 2023 at 20 families, then eased off — this looks like a claim-staking wave that has already crested rather than an accelerating land grab.
- One applicant pairing anchors the field: Qualcomm filing jointly with named inventors Li Qiaoyu and Taherzadeh Boroujeni Mahmoud accounts for the strongest co-assignee links in the dataset.
- Claims cluster tightly in wireless architecture, with H04W and H04B covering the great majority of records while G06N-classified AI-model claims stay under 4% of the corpus.
What this patent landscape covers
This landscape tracks patent families at the intersection of non-terrestrial networks — satellite, HAPS and other space-based access layers — and the machine-learning techniques now being claimed to manage them: predictive beam management, AI-based handover decisions, and ML-driven resource allocation. The search spans filings from 2015 through the 2026 data cut-off, drawing on 55 published records treated as 55 distinct families for ranking purposes.
Because publication typically lags filing by around 18 months, the 2025 and 2026 counts in any trend line are still filling in and should be read as provisional rather than as a genuine falloff.
Filing trend and technology composition
Two views of the same 55 families: how filing activity has moved year over year, and which IPC subclasses the claims actually sit in.
A wave that crested in 2023
Filings moved from zero in 2017 to a peak of 20 in 2023, with the 2022 midpoint at 12. That trajectory is flat-to-declining rather than compounding, which is unusual for a topic this current — it suggests an initial burst of foundational filing around predictive beam management that has not yet been followed by a second wave.
Wireless architecture dominates the claim space
H04W (wireless communication networks) appears in 38 of the records and H04B (transmission, general) in 21, together covering the bulk of the corpus. H04L (digital transmission) trails at 14. Genuinely AI-classified claims under G06N sit at just 2 records and G06Q business-process claims at 1 — the machine-learning element is overwhelmingly being claimed as a wireless-system feature, not as a standalone AI method.
Shares are the percentage of the 55 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Non-Terrestrial Network AI and Machine Learning with Eureka
This page is one run against one query. Ask Eureka your own question about non-terrestrial network ai and machine learning and every answer comes back with the patent numbers behind it.
Try EurekaThe most-cited and most recent filings
L1 reporting enhancement in mTRP for predictive beam management
This disclosure provides systems, devices, apparatus, and methods, including computer programs encoded on storage media, for reporting enhancements to predictive beam management. A UE may measure at least one of an RSRP or a SINR associated with one or more secondary CMR-IDs of a CMR set, distinct from the primary CMR-ID, and report those measurements based on one or more CMR-ID selection protocols.Filed by Qualcomm; published 2025-06-05.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | WO2024007248A1 | Layer 1 report enhancement for base station aided beam pair prediction | 16 |
| 2 | WO2023149713A1 | Method and apparatus for configurable measurement resources and reporting in wireless communication system | 3 |
| 3 | US20240188025A1 | Signaling between a perception-assistance node, a user equipment (UE), and a base station | 3 |
| 4 | WO2024207416A1 | Inference data similarity feedback for machine learning model performance monitoring in beam prediction | 2 |
| 5 | WO2024207182A1 | Training dataset mixture for user equipment-based model training in predictive beam management | 2 |
| 6 | US20250330855A1 | Layer 1 report enhancement for base station aided beam pair prediction | 2 |
| 7 | US20260214532A1 | Timed device handover event reporting | 1 |
| 8 | WO2025236963A1 | 通信方法、通信装置和系统 | 1 |
| 9 | WO2023206121A1 | L1 reporting enhancement in MTRP for predictive beam management | 1 |
Citation counts favour older filings simply because they have had longer to be cited; treat this as a signal of influence within the searched corpus, not a ranking of current importance.
Patent titles are shown in the language they were filed in, not translated, so that each record stays verifiable against the original filing — a translated title will not match in Eureka or in any national register. Publication numbers are shown where the record carries one (9 of 9 rows); clicking a row searches Eureka by that number.
Put your own technology through the same analysis
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 →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 →What the numbers say about this field
Three patterns stand out once the filing trend, IPC mix and citation data are read together.
The peak has already passed
Filings rose from zero in 2017 to 20 in 2023, then did not keep climbing — 2022's midpoint of 12 sits well below the peak, and the flat-to-declining shape suggests the founding claims around predictive beam management and AI-based handover are largely staked. New entrants now face denser prior art than the raw novelty of the topic would suggest.
One filing cluster dominates co-ownership
The strongest co-assignee link in the dataset — Qualcomm paired with named inventors Li Qiaoyu and Taherzadeh Boroujeni Mahmoud at 15 shared records — plus a second pairing at 12, points to a single research group driving a disproportionate share of the claim activity around beam prediction and reporting enhancements.
Claims sit in wireless systems, not AI methods
H04W and H04B together account for the great majority of records, while G06N — the subclass for AI-model computing itself — covers only 2 records. Applicants are overwhelmingly claiming machine learning as an embedded feature of a wireless protocol (measurement, reporting, handover) rather than as a general AI technique.
PCT is the primary entry route
Of the recorded receiving offices, WIPO/PCT accounts for 24 filings, ahead of Europe at 13, India at 8, the United States at 7 and China at 3. A PCT-first strategy dominates, consistent with applicants seeking broad optionality before committing to national phase in specific markets.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to non-terrestrial network ai and machine learning, with the prior art for and against each one.
Who is filing, and where the gate sits
Filing activity concentrates around a small cluster of related applicants and named inventors, with recent-year momentum mostly flat.
Qualcomm holds the most active recent filing rate
Qualcomm is the only assignee in the recent-momentum data showing any activity in the latest year — 3 filings, flat year-on-year — while every other tracked assignee, including the named inventors most closely tied to Qualcomm's filings, shows zero in the same period.
A tight inventor-assignee cluster
Li Qiaoyu and Taherzadeh Boroujeni Mahmoud each co-file heavily with the same corporate assignee, and with each other, forming the strongest links among the 10 co-assignee pairs identified. This is the signature of a dedicated internal research team rather than a broad industry consortium.
Other named parties have gone quiet
Pezeshki Hamed, Samsung Electronics, Luo Tao and others appear in the ranking but show no filings in the latest tracked year. Whether that reflects publication lag or a genuine pause cannot be separated from this data alone, but it leaves the recent pipeline visibly dependent on a single applicant.
| Assignee | Recent year | YoY |
|---|---|---|
| Qualcomm Incorporated | 3 | 0% |
| LI QIAOYU | 0 | — |
| TAHERZADEH BOROUJENI MAHMOUD | 0 | — |
| PEZESHKI HAMED | 0 | — |
| Samsung Electronics Co., Ltd. | 0 | — |
| LUO TAO | 0 | — |
| Huawei Technologies Co., Ltd. | 0 | -100% |
| DAMNJANOVIC JELENA | 0 | — |
Where to take this research
The filing data points to a narrow, concentrated field with a specific applicant driving recent activity — worth validating against your own freedom-to-operate position.
Map claims against your own architecture
Check whether your beam-management, handover or resource-allocation implementation reads on the Qualcomm-anchored claim cluster before committing engineering resources.
Explore in EurekaWatch for the next filing wave
The 2023 peak has not been followed by renewed growth in the visible data, but publication lag means 2025–2026 filings are still arriving — track this quarterly rather than assuming the field has settled.
Set up monitoring in EurekaProbe the under-claimed branches
G06N-classified AI-model claims and cross-satellite handover coordination show low density relative to core wireless claims — worth a deeper prior-art check before assuming the space is open.
Run a white-space search in EurekaCommon questions about this patent landscape
Within this dataset, Qualcomm is the most active recent filer, showing 3 filings in the latest tracked year while other tracked assignees show none. The strongest co-assignee relationships also centre on Qualcomm, filing jointly with named inventors Li Qiaoyu and Taherzadeh Boroujeni Mahmoud at 15 and 12 shared records respectively. That concentration suggests a single internal research group is currently the most influential filer in this specific claim space, though the broader satellite-communications field has many other participants outside this narrower AI/ML search.
Not on the visible trend: filings rose from zero in 2017 to a peak of 20 in 2023, then eased rather than continued climbing, with the 2022 midpoint at 12. Because publication lags filing by roughly 18 months, the 2025 and 2026 figures are undercounted and should not yet be read as a real decline. A fair read is that the founding wave of claims has been filed and the field is waiting to see whether a second wave follows.
Overwhelmingly wireless protocols. H04W (wireless communication networks) and H04B (transmission, general) cover the large majority of the 55 records, while G06N, the subclass for AI-model computing itself, appears in only 2 records. In practice this means most claims describe machine learning as an embedded feature — beam prediction, measurement reporting, handover triggers — rather than claiming a novel AI or ML method in the abstract.
The IPC composition points to G06N-classified AI-model claims and G06Q business-process claims as thin relative to the dominant H04W/H04B wireless-architecture filings — only 2 and 1 records respectively out of 55. Cross-satellite handover coordination and on-board inference resource scheduling are adjacent branches that show similarly low density in this corpus. Any of these would need a dedicated prior-art search beyond this dataset before relying on the gap, since low density here reflects this specific search string, not the whole patent universe.
Not necessarily. Citation counts accumulate over time within a searched corpus, so older filings such as WO2024007248A1 on layer-1 report enhancement for beam pair prediction naturally collect more citations than a 2025 filing simply by having been public longer. Treat citation rank as a measure of historical influence on later filers, and cross-check it against recency and claim scope before deciding a document is commercially significant now.
Research Non-Terrestrial Network AI and Machine Learning in depth with Eureka
Go past this page: query the whole non-terrestrial network ai and machine learning 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.