Mixture of Experts Patents: Top Companies & Filing Trends 2026
Mixture of Experts Training Patents: Who Is Filing and Where the Claims Sit
+850% filing growth from 2021 (2 records) to 2024 (19 records), the last year the data can treat as complete. Filings moved from 0 in 2017 to 18 in the most recent year, with a peak of 46 in 2025. The 2021-to-2024 span alone shows +850% growth (2 to 19 records) — the clearest complete-year signal in the set, since publication lag means 2025 and 2026 will still fill in.
- 1HUAWEI TECH CO LTD18
- 2MICROSOFT TECHNOLOGY LICENSING LLC16
- 3UNIFABRIX LTD10
- 4SIEMENS HEALTHINEERS INTERNATIONAL AG7
- 5BEIJING BAIDU NETCOM SCI & TECH CO LTD5
See the full mixture of experts 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 about this landscape
Who holds the most mixture of experts training patents?
One assignee leads the ranked list with 18 of the 108 records in scope, well ahead of fifth place at 5 records and tenth place at 2. The ranking covers 49 companies in total, and the drop-off from the leader to the rest of the field is steep, meaning most other filers have only one or two records rather than a sustained programme. This concentration pattern — one clear leader plus a long tail — is typical of a technology area that is still young but where one organisation moved early.
How fast is patent filing growing in mixture of experts training?
Using the last full year of complete data, filings grew from 2 records in 2021 to 19 in 2024, a rise the dataset states as +850% over that span. The two most recent years in the trend, 2025 and 2026, will continue to fill in as more applications publish, since publication typically lags filing by around 18 months. Treat the 2024 figure as the most reliable recent read on growth, not the years after it.
What patent classes cover mixture of experts training?
The two dominant IPC subclasses are G06N (computing arrangements based on specific computational models, i.e. AI) at 59.3% of the 108 records, and G06F (electric digital data processing) at 57.4%. Smaller shares appear in image processing (G06T, G06V) and digital transmission (H04L), each at 5.6%. Because a single record can carry multiple IPC classes, these shares add up to more than 100%, and that is expected rather than an error.
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 Mixture of Experts 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.