OPC & Computational Lithography Patents: Leaders & Trends 2026
- Filing has flattened, not grown. annual filings peaked at 72 in 2022 and have since declined toward single digits in the partial 2026 count — a maturing claim space, not an expanding one.
- The US and China dominate filing location. 403 US and 212 China receiving-office records together account for the large majority of the 806 families tracked, with Europe, Taiwan and Japan trailing well behind.
- Momentum has gone quiet across nearly every major filer. recent-year filings from ASML, Samsung and SMIC show flat or -100% YoY activity, with only TSMC and Shanghai Huali showing any residual filing.
Filing growth compares 2021 (38 records) with 2024 (55) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 806 records in scope (CR5), not by the ranked leaders only.
What this patent set covers
Optical proximity correction (OPC) and computational lithography cover the software and modelling techniques used to pre-distort mask patterns so that printed features on a wafer match the intended design despite diffraction and process variation. This landscape tracks 806 patent families filed between 2015 and mid-2026 that combine OPC, computational lithography or source mask optimization claims with process-window, model-calibration, mask-rule-check, runtime or inverse-lithography language in their title or claims — a filter designed to isolate substantive method and system claims rather than incidental mentions.
The IPC composition confirms the field's centre of gravity: G03F (photolithography and photomechanics) appears in 784 of the 806 records, with G06F (digital data processing) present in 258 — evidence that this is fundamentally a computational discipline layered onto a physical patterning process, not a pure optics field.
Trend and technology composition
Two views of the same dataset: how filing volume has moved year over year, and which IPC subclasses the claims actually sit in.
A peak in 2022, then decline
Filings rose from 27 in 2017 to a peak of 72 in 2022, then declined toward 7 in the most recent (partial) year. Because publication typically lags filing by around 18 months, the last one to two years understate true filing activity — but the pre-2022 trajectory and the post-peak drop are both large enough to read as a genuine plateau rather than a data artefact.
G03F carries the claims; G06F and H01L support them
G03F (photolithography) covers 784 of 806 records, confirming this is squarely a lithography-process filing set. G06F (258) and G06N (44, AI-based computing) show the computational and machine-learning layer growing inside that core, while H01L (119) reflects downstream semiconductor-device claims tied to the corrected mask output.
Shares are the percentage of the 806 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Optical Proximity Correction and Computational Lithography with Eureka
This page is one run against one query. Ask Eureka your own question about optical proximity correction and computational lithography and every answer comes back with the patent numbers behind it.
Try EurekaThe most-cited foundations in this set
Optical proximity correction based on combining inverse lithography technology with pattern classification
A layout design is received, then a machine-learning-based clustering process separates layout features into groups. Preliminary corrections are determined for each group using inverse lithography technology, applied to produce a pre-processed layout, and a further inverse lithography technology process refines the result.Filed by Siemens Industry Software, published 2023-03-30 — illustrates how pattern classification is now used to reduce the search space before inverse lithography runs, targeting the runtime problem that has dogged full-chip ILT for years.
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US7175940B2 | Method of two dimensional feature model calibration and optimization | 367 |
| 2 | US6194104B1 | Optical proximity correction (OPC) method for improving lithography process window | 329 |
| 3 | US20080309897A1 | Multivariable solver for optical proximity correction | 313 |
| 4 | US6578190B2 | Process window based optical proximity correction of lithographic images | 297 |
| 5 | US6749972B2 | Optical proximity correction common process window maximization over varying feature pitch | 218 |
| 6 | US6426269B1 | Dummy feature reduction using optical proximity effect correction | 211 |
| 7 | US7123356B1 | Methods and systems for inspecting reticles using aerial imaging and die-to-database detection | 138 |
| 8 | US6881523B2 | Optical proximity correction method utilizing ruled ladder bars as sub-resolution assist features | 121 |
| 9 | US20030082463A1 | Method of two dimensional feature model calibration and optimization | 110 |
| 10 | US9588438B2 | Optimization flows of source, mask and projection optics | 96 |
Citation counts favour older filings by construction — a patent published in 2000 has had two decades longer to accumulate citations than one published in 2023. Read this table as a map of foundational method claims, not as a ranking of current relevance.
Each row carries its publication number; 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 direction
Three read-outs from the filing and citation data that matter for anyone deciding where to file next.
Volume has plateaued since 2022
The filing count rose steadily through the late 2010s, peaked at 72 families in 2022, and has fallen since. Some of the recent drop is the normal 18-month publication lag, but the shape of the decline starts well before the most recent year, suggesting genuine cooling in new filing activity rather than a reporting gap alone.
US and China together hold the bulk of filings
Of 806 families, 403 were filed at the US receiving office and 212 at China's, with WIPO, Europe, Taiwan and Japan each taking a smaller share. That split reflects where the fabs and EDA vendors that write these claims are actually headquartered or litigating.
The oldest core methods still anchor the field
The five most-cited records in this set are all pre-2010 US filings covering model calibration, process-window optimization and multivariable OPC solvers. Their citation counts are a legacy effect of age, but the claim language they established — process-window-based correction, common-window maximization — still frames how newer filings describe their improvements.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to optical proximity correction and computational lithography, with the prior art for and against each one.
Who is still active, and who has gone quiet
Recent-year momentum is a better signal than cumulative filing counts here, because it shows who is still investing versus who is coasting on an established portfolio.
Two of the historically largest filers show zero recent activity
ASML and Samsung both show 0 filings in the latest year against a prior year of activity, a -100% YoY reading. That does not mean these firms have exited the field — publication lag means recent internal filings may not yet be visible — but it does mean their public claim activity has gone quiet for now.
Foundries are the only segment still filing, thinly
TSMC shows 1 filing in the latest year, down 50% YoY, and Shanghai Huali shows 1 filing, flat YoY. Both are foundries with direct manufacturing exposure to OPC accuracy, which likely explains why they are the last segment still adding to the record while EDA and equipment vendors have paused.
SMIC's Shanghai and Beijing entities file jointly at scale
The strongest co-assignee pair in the dataset links SMIC's Shanghai and Beijing manufacturing entities across 15 shared families, far ahead of the next pairs (ASML's Netherlands and holding entities at 4; D2S and its named inventor at 4). That concentration suggests centralized IP strategy across SMIC's fab sites rather than independent local filing.
| Assignee | Recent year | YoY |
|---|---|---|
| Taiwan Semiconductor Manufacturing Company (TSMC) | 1 | -50% |
| Shanghai Huali Microelectronics Corporation | 1 | 0% |
| ASML Netherlands B.V. | 0 | -100% |
| Samsung Electronics Co., Ltd. (South Korea) | 0 | -100% |
| Semiconductor Manufacturing International (Shanghai) Corporation (SMIC) | 0 | — |
| ASML Mask Tools B.V. | 0 | — |
| D2S, Inc. | 0 | — |
| United Microelectronics Corporation (UMC) | 0 | — |
Where to take this analysis
The dataset points to a field with occupied core claims and a few specific gaps. Deciding whether to file, license or design around requires going deeper than this summary.
Run a freedom-to-operate check on the gate chips
The under-claimed branches identified here — runtime reduction for full-chip ILT, EUV-specific stochastic modelling — are candidates for new filing, but confirming they are genuinely open requires a full claim-by-claim search, not just a family count.
Explore in EurekaTrack the quiet majors for re-emergence
ASML and Samsung show -100% YoY momentum, which given publication lag could mean filings already exist but have not published yet. Set up ongoing monitoring rather than treating the quiet period as final.
Explore in EurekaMap the foundational citation cluster against your own claims
The top five cited patents anchor process-window and multivariable-solver language used across the field. Any new filing in this space should be checked against that cluster specifically, not just against the most recent filings.
Explore in EurekaCommon questions about this landscape
A patent family groups all the related filings — across jurisdictions and continuations — that trace back to the same original invention. This landscape counts 806 such families rather than raw published documents, because a single invention filed in the US, China and Europe would otherwise be counted three or more times. Using families gives a fairer picture of how many distinct inventions exist, rather than how aggressively any one applicant pursued multi-jurisdiction or continuation filing.
Filings rose from 27 in 2017 to a peak of 72 in 2022 before falling toward single digits in the most recent partial year. Part of that decline is a publication-lag artefact — patents typically publish around 18 months after filing, so the last one to two years are always undercounted at the point of any given data pull. But the drop begins visibly before the most recent year, which suggests real cooling in new filing activity, possibly reflecting that core OPC correction methods are now well covered by existing claims, pushing new work toward adjacent, less-claimed branches.
Recent-year momentum data shows the field has gone quiet across most historically active filers: ASML and Samsung both show 0 filings in the latest year, a -100% year-over-year change. TSMC and Shanghai Huali show minimal but nonzero activity — one filing each in the latest year — making foundries with direct manufacturing exposure the last visibly active segment. This does not necessarily mean equipment and EDA vendors have stopped innovating; it may reflect filings not yet published.
The IPC composition and gate analysis point to several under-claimed branches relative to the dense core OPC correction claim space: runtime reduction techniques for full-chip inverse lithography, EUV-specific stochastic OPC modelling, and machine-learning-driven mask rule check automation. These areas show thinner claim density in this dataset, but that should be confirmed with a dedicated freedom-to-operate search before treating them as genuinely open, since thin representation in one search string is not proof of an empty field.
Of 806 tracked families, 403 were filed at the US receiving office compared with 212 for China, and smaller counts for WIPO, Europe, Taiwan and Japan. This split reflects where the major EDA software vendors, equipment makers and litigation activity in lithography have historically concentrated, alongside China's growing foundry sector filing domestically to protect its own manufacturing processes. It does not mean the underlying technology is only relevant in those two markets — just that those are where applicants have chosen to seek protection first.
Research Optical Proximity Correction and Computational Lithography in depth with Eureka
Go past this page: query the whole optical proximity correction and computational lithography 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.