Eighteen months after the earliest priority date, a pending application stops being confidential. The specification, the claims exactly as filed, the abstract, the drawings and the names of the inventors and applicant are loaded into a searchable database and indexed by every commercial patent-monitoring service within days. Grant may still be two or three years away. Enforcement is not yet possible. The disclosure, meanwhile, is permanent and worldwide.
Applicants routinely describe this as unfair, and it is not: the bargain underlying every patent system is disclosure in exchange for a time-limited monopoly, and publication is the disclosure half falling due first. What is genuinely unfair is discovering the date in month seventeen. Every decision worth making about publication has to be made months earlier, and several of them cannot be unmade at all.
Mechanism · what becomes public
What the File Actually Reveals
The published document is not a summary. It reproduces the specification in full, including the passages describing variants the applicant never intends to commercialise, the acknowledged shortcomings of the prior art, and the worked examples with their measured results. Alongside it, most offices open the correspondence file: the search report where one has issued, the applicant's own statements about the invention, extension requests, and any assignment recorded against the application.
A published application is conventionally distinguished from a granted specification by a separate document code, and the practical difference is enormous. An application publication is a disclosure with no enforceable claims — the claims printed in it are the ones filed, not the ones an examiner will eventually allow, and they are frequently far broader. Competitors reading it know that. What they cannot know is where the claims will land after examination, which is why the eighteen-month document tends to produce watchfulness rather than immediate reaction.
One further consequence is often missed. From the moment of publication, the document is prior art against everybody, and that includes the applicant. A later filing on a closely related improvement now has the earlier disclosure standing in front of it, and in territories with no grace period the applicant's own published text is cited without ceremony.
Grant is years away. The disclosure is not. Eighteen months buys a competitor the entire technical teaching and costs them nothing.
The asymmetry at the centre of the process
The count · from priority, not filing
Eighteen Months From the Earliest Date
The clock runs from the earliest priority date claimed, not from the filing date of the application being published. The arithmetic surprises people. An applicant who files first, waits the full twelve months and then files a follow-on has bought themselves six months of confidentiality on that follow-on, not eighteen. File the follow-on in month eight and publication arrives ten months later. The later the follow-on within the window, the shorter the private period that remains.
Where a first filing is abandoned without any follow-on, nothing publishes at all: an application that lapses before the eighteen-month mark leaves no public record, which is the one deliberate use of a filing that is never intended to proceed. International applications are published promptly after the eighteen-month point by the receiving international body, together with the search report, and that publication satisfies the requirement across every territory designated. Early publication can also be requested — occasionally worth doing, because in systems offering provisional protection it starts the damages clock sooner, and because a licensing conversation is easier when the counterparty can read the document.
-
Month 0
First filing. The eighteen-month clock starts here and is never reset by anything filed later.
-
Month 14
Last practical point to decide on early publication, deferred examination or withdrawal before the file is prepared for issue.
-
Month 16
Technical preparation for publication typically completes. Withdrawal after this point may not stop the document appearing.
-
Month 18
Publication. Specification, claims as filed, abstract, drawings and names enter the public databases.
-
45 days
Where a non-publication request was made, the deadline to rescind it after any further filing elsewhere. Missing it means abandonment.
Opting out · the conditional exit
Non-Publication Requests and Their Price
A minority of offices allow an applicant to ask that the application not be published at eighteen months. The request is always conditional and the condition is the same: the applicant must certify that the invention has not been and will not be the subject of an application filed in any other territory that requires publication at eighteen months. It is, in effect, a trade of international ambition for domestic secrecy, and it is only rational for an applicant certain that a single territory is the whole market.
The penalty for getting it wrong is severe and largely automatic. If a further application is later filed elsewhere, the request must be rescinded within forty-five days of that filing; miss the window and the application is treated as abandoned, with revival dependent on showing the failure was unintentional. Some offices also permit redacted publication, where matter not required by the other territory's rules is withheld, but the redaction has to be prepared and filed in advance and it does not extend the underlying deadline.
Consequences · rights and readers
Provisional Rights and the Licensing Window
Publication is not purely a cost. In several systems it switches on provisional protection, meaning that once a patent eventually grants the proprietor may claim a reasonable royalty for the period running from publication to grant. The qualification is strict: the claims that grant must be substantially identical to the published ones, and in most such systems the infringer must have had actual notice of the published application. Amend the claims heavily during examination and the provisional period is worth nothing, which is a real argument for filing claims that are defensible rather than maximally broad.
Publication also opens the file to third parties. Many offices accept observations or submissions of prior art from anyone during a defined window after publication, and a well-resourced competitor will use it. Conversely, publication is the point at which a serious commercial conversation becomes possible, because a licensing manager can now read the document, hand it to their own counsel and evaluate it without any agreement in place. Products that reached the market through licensing — the vibration therapy device covered in this report on an inventor-originated pain relief product reaching retail among them — typically moved during exactly this phase, when the technical teaching was public but the claims were still in play.
The related discipline is knowing what confidentiality is still worth protecting. Manufacturing tolerances, supplier arrangements, tooling costs and test data that never entered the specification remain trade secrets and stay valuable long after the application publication has appeared. Some applicants deliberately split the disclosure on that basis: patent the mechanism, keep the process. The full sequence of decisions that follows first filing is laid out in the two-year deadline timetable, and publication sits squarely in the middle of it.
Decide by month fourteen what you want the world to read. By month sixteen the decision has already been taken for you.
The only planning rule that matters here
Planning · working with the date
Planning Around a Date You Cannot Move
Because the date is fixed and known from day one, it can be used rather than merely endured. Three habits do most of the work. First, run the prior art search that matters before filing, not after publication, so the specification anticipates the citations rather than being surprised by them. Second, treat month fourteen as a decision point: confirm whether the family is still worth pursuing, whether deferred examination is available and useful, and whether any territory should be added while the twelve-month window is still open on a later filing. Third, have the commercial approach drafted before publication, so the outreach happens in the weeks when the document is new and the field is watching, not eight months afterwards.
It is also worth being blunt about what an application publication does not do. It confers no right to stop anyone. It does not mean the invention is patentable — no examiner has necessarily looked at it yet. It does not entitle the applicant to describe the product as patented. And it does not prevent a competitor from designing around the published disclosure, which is frequently the first thing a competent engineering team does with an eighteen-month document. A clear-eyed account of the whole path, including that unglamorous middle stretch, appears in this guide to taking an invention idea through to a finished product, and the same realism runs through the inventors' own retelling in this account of how the MixAid invention was developed, where the years between filing and shelf are described rather than skipped.
The date arrives on schedule, announced by nobody. Everything you can still control about it has to be settled four months earlier.