Structured and Unstructured Pruning Patents: Who Leads and Where Filings Concentrate
71.5% concentration. The top 5 assignees hold 435 of 608 records in scope — a field with a small set of firms controlling most of the claim space. Annual filings rose from 5 records in 2017 to a peak of 146 in 2022, then settled into a higher plateau — 67 records in 2021 against 73 in 2024, a 9% increase over that three-year span. Because publication lags filing by around 18 months, the 2025-2026 figures shown are still filling in and should not be read as a slowdown.
- 1INTEL CORP373
- 2NVIDIA CORP25
- 3MITSUBISHI ELECTRIC CORP14
- 4HUAWEI TECH CO LTD12
- 5GOOGLE LLC11
See the full structured and unstructured pruning 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 pruning patents
Who holds the most patents in structured and unstructured pruning?
One assignee leads the field with 373 of the 608 records in scope, and the top 5 assignees combined hold 435 records — 71.5% of the total. The top 10 extend that to 487 records, or 80.1% of the field. This means the field is concentrated at the top with a long tail of smaller filers below the leading group, rather than evenly spread across many competitors.
Is patent filing in neural network pruning still growing?
Filing rose sharply through the late 2010s and peaked at 146 records in 2022, then settled onto a plateau; 2021 recorded 67 filings against 73 in 2024, a 9% increase over that span. Because publication typically lags filing by around 18 months, the 2025 and 2026 figures in any trend view are still incomplete and should not be read as evidence of a slowdown. 2024 is the most recent year that can reasonably be treated as a complete picture.
What technology areas do pruning patents cover most heavily?
Digital data processing (IPC class G06F) appears on 51.5% of the 608 records, followed by AI model computing (G06N) at 37.3% and image data processing (G06T) at 33.4%. Hardware-facing classes such as coding and code conversion (H03M, 3.1%) and non-electric control systems (G05D, 1.5%) are much thinner, indicating that training-side and masking mechanics are more heavily claimed than execution hardware. Since a single record can carry multiple IPC codes, these percentages sum to more than 100% and each should be read against the same 608-record base.
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 Structured and Unstructured Pruning 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.