Graph Neural Network Training Patents: Leaders & White Space 2026
Graph Neural Network Training Patents: Who Is Filing and Where the Field Is Still Open
No single owner dominates. The leader holds 151 records out of 2,070, and the top 5 combined reach only 13.1% of all records in scope — filing is spread across a long tail rather than locked up. Filings moved from 0 in 2017 to a peak of 612 in 2025, with 2026 sitting at 171 so far. The 2021-to-2024 span — the last stretch of years complete enough to trust — shows filings rising from 200 to 387, a 94% increase.
- 1ADOBE INC7.3%
- 2TENCENT TECHNOLOGY (SHENZHEN) CO LTD2.1%
- 3HUAWEI TECH CO LTD1.4%
- 4TATA CONSULTANCY SERVICES LTD1.2%
- 5BEIJING UNIV OF POSTS & TELECOMM1.2%
See the full graph neural network training analysis in Eureka
- The complete ranking, not just the top five
- Every IPC branch with its share of the corpus
- The most-cited records, and where claim space is still thin
Common questions on graph neural network training patents
How many patents exist for graph neural network training?
This dataset tracks 2,070 published records filed between 2015 and mid-2026 that combine core graph neural network or embedding terms with training-specific techniques such as subgraph sampling, mini-batch training and heterogeneous graph aggregation. Filing was negligible before 2018 and grew steadily afterward. Because publication lags filing by roughly 18 months, the most recent one to two years in any such count will keep rising as more applications publish.
Who are the leading patent filers in graph neural network training?
The leading assignee in the ranked dataset holds 151 records, well ahead of the fifth-place holder at 24 and tenth place at 17, so there is a clear single leader followed by a steep drop-off. The top 5 assignees combined account for 13.1% of all 2,070 records in scope, and the top 10 combined reach 18.0% — meaning most filing activity sits outside the leading group entirely. The ranked list spans 100 companies including technology vendors and universities across China, the US and India.
Is the graph neural network training patent field growing or slowing down?
Growing, based on complete-year data: filings rose from 200 in 2021 to 387 in 2024, a 94% increase over three years. The apparent decline visible in 2025 and 2026 figures is a publication-lag artefact rather than a real slowdown, since patent applications typically take about 18 months to publish. 2025 is currently the peak year on record at 612 publications, and that number will likely rise further as more filings from that year become visible.
Disclaimer. This analysis is based on Patsnap Eureka data drawn from a limited snapshot of global patent records and is provided for general information and reference only. Patent data carries inherent limitations — recent filings are under-counted because of publication lag, counts may be on a record or family basis, classification and applicant-name data may contain errors or duplicates, and the underlying search query defines the scope shown — so the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing here is an exhaustive prior-art, novelty, freedom-to-operate or validity search, nor does it constitute legal, financial or professional advice, and it should not be relied upon as such. Verify independently and review with qualified patent and legal professionals before acting on it.
Method: Filing trend and technology composition. Derived from a Patsnap search on Graph Neural Network Training covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish. Every share divides by all records in scope. Data: Patsnap Eureka. See the full landscape report.