What Happens During Patent Prosecution

The short version

Three in four publicly observable applications had issued as patents. Of 1,308,614 US utility applications filed between 2020 and 2022 that are visible in the public record, 76.3% had been granted by 26 October 2026, 19.0% had ended without a patent and 4.8% were still pending.

Publicly observable matters here. An application abandoned before publication never becomes public, so it is absent from this count. About 3.6% of serial numbers from the period are missing for that reason, which would put the true per-application rate no lower than about 73.9% if every one of them had failed.

Most accounts of prosecution are written from experience. These five pages are written from the record, and in several places the record disagrees with the folklore.

Prosecution is the part of patent practice people describe to each other in anecdotes. How long it takes, how many rejections to expect, whether a final rejection is the end, what a request for continued examination actually buys. The anecdotes are usually about the memorable cases rather than the typical ones.

These pages answer five questions from the USPTO's own records, each with its population, its period and its limits printed alongside. They are counts of what happened. They are not predictions, and they are not advice about any application.

How applications filed 2020 to 2022 ended

Status of 1,308,614 publicly observable US utility applications
Filed 2020 to 2022, status as of 26 October 2026, not an eventual disposition. Bars run from zero.
Granted
76.3%
Ended without a patent
19.0%
Still pending
4.8%

The pending share will mostly resolve into the first two. Counted per family rather than per application the grant share is 75.0%, because application-level counting gives more weight to families that filed several applications.

Three in four is also a recent number rather than a standing fact. Across earlier cohorts the share granted fell from 73.8% of 2001 filings to a trough of 61.7% for 2006 filings, then rose every year after that to 79.8% for 2019. Anyone who formed an impression of the odds from the mid-2000s formed it at the bottom of that curve.

It also depends on what you count as success. Of the same 899,034 originals, 58.8% were granted without any continuing application being filed from them, against 75.0% granted by any route. Treating a grant that needed a continuation as a failure is a defensible definition and it moves the answer by 16 points, which is worth knowing before quoting any single grant rate.

Each page below uses the population that fits its own question, so the cohorts differ: outcomes here are 2020 to 2022 filings, timing uses the 2021 cohort, rejection grounds come from Office actions mailed 2024 to 2025. They are related questions rather than one funnel, and percentages from different pages should not be multiplied together.

The five questions

How long does a patent take? About two to three years for the typical filer, and far longer in biotechnology. Measured over an equal window, grants got slower rather than faster. Most of the pendency is a single stage. Read the Answer How many rejections? Half of granted applications saw exactly one rejection before allowance and one in five saw none. What the rejections are about, and which grounds come back in the next action. Read the Answer After a final rejection Three in five continue with a request for continued examination. Seven in ten are granted within four years, one RCE is the commonest route, and almost nothing is still open at four years. Read the Answer How applications end Abandonment is nearly always silence: a reply period runs out. Express withdrawal is under two percent. Whether the family continues afterwards, and why it usually does not. Read the Answer Does who files change it? Small and micro entities are granted less often than large ones. More than half of that gap goes with technology mix and who files the application, and a gap remains after both. Examiners in the same art unit also differ. Read the Answer

What these pages do not do

These are descriptions, not forecasts. Nothing here tells a reader what their application will do, what they should file, or when. Every figure is an association drawn across large populations, and the things that most affect a particular application, the claims and the art, are not in these counts at all.

Our own counts from the USPTO record. These describe what happened across past applications. They say nothing about any particular application, examiner or firm, they are associations rather than causes, and none of this is legal advice.

The figures on this page are available as a CSV: /data/what-happens-during-prosecution.csv. It is generated from the tables and charts above, so it carries the same numbers and cannot fall out of step with them.

Related

Patent Continuations A continuation turns one specification into a patent family: new claims, same priority date, no new matter. Why disclosed-but-unclaimed scope can be dedicated to the public, the copendency and term rules, the double-patenting tradeoff, and an honest read on where AI drafting helps. Read the Guide Filing and Fees Four changes to what a filing costs and how it must be made. No thresholds or qualification criteria, on purpose. Read the Cluster When Must I File? The deadline arrives inside an ordinary commercial decision, before most companies think patents are relevant. Read the Cluster