A patent invalidity search looks backwards from a specific patent, hunting for anything published before its priority date that reads on its claims.
The target is known, which changes everything. A patentability search asks an open question and can stop at the first good reference. An invalidity search has the claims in front of it and must be thorough enough to rely on.
Which is why it costs more. Typically $5,000 to $20,000, against $500 to $3,000 for a patentability search, and the difference is exhaustiveness rather than skill.
The best references are usually not patents. Examiners search patents well and everything else less consistently, so the gap is where invalidity searches produce results.
How it differs from a patentability search
| Patentability search | Invalidity search | |
|---|---|---|
| Question | Can I patent this? | Should this patent have issued? |
| Target | An invention concept | Specific issued claims |
| Can stop early | Yes — one good reference answers it | No — must be thorough |
| Date cutoff | Today | The patent's priority date |
| Typical cost | $500–$3,000 | $5,000–$20,000 |
| Used for | Filing decisions | Defence, PTAB petitions, diligence |
Having the claims is an advantage. The search can be aimed element by element rather than at a general concept, which makes it far more precise.
Thoroughness is the cost driver. A patentability search that misses something is a filing risk; an invalidity search that misses something loses a case.
The priority date governs everything
| Date | Relevance |
|---|---|
| Earliest effective priority date | The cutoff — art must predate it |
| Provisional filing date | Often the earliest, if properly supported |
| Parent application filing date | For continuations and divisionals |
| Foreign priority date | Counts for prior art purposes |
| Patent filing date | Frequently much later — do not use it |
| Grant date | Irrelevant to the search |
Establishing the correct date is step one and it is easy to get wrong. A patent filed in 2019 as a continuation of a 2013 application has a 2013 cutoff, and a search run from 2019 will return six years of irrelevant material while missing the window that matters.
Priority is only as good as the disclosure supporting it. A claim not enabled by the provisional does not get the provisional's date, which occasionally opens a window of art the patent owner assumed was excluded.
Check the front page and the file history together. Related application data gives the chain; the file history shows whether priority was ever challenged.
Where the best references hide
| Source | Examiner coverage | Value to an invalidity search |
|---|---|---|
| US patents | Thorough | Low — already considered |
| Foreign patents | Moderate | Good |
| Conference proceedings | Weak | High |
| Trade publications | Weak | High |
| Standards documents | Weak | High |
| Technical manuals and datasheets | Very weak | High |
| Theses and dissertations | Very weak | Good |
| Archived websites | Very weak | Good, with date evidence |
The pattern is consistent: the less thoroughly examiners search a source, the more valuable it is. US patents are the most searched category and therefore the least likely to yield something new.
Non-patent literature is where invalidity searches earn their fee. It is also what separates a $2,000 search from a $12,000 one.
Date evidence matters for non-patent sources. A conference paper needs a demonstrable publication or presentation date, and an archived web page needs archive evidence.
The knockout search first
| Knockout search | Full study | |
|---|---|---|
| Cost | Fraction | Full |
| Time | Days | 2–4 weeks |
| Scope | Obvious candidates | Exhaustive |
| Purpose | Is there anything obvious? | Build the case |
Run the cheap version first. A knockout search occasionally finds something decisive and ends the question for a small fraction of the full cost.
It also informs the brief. A knockout that finds nothing in patents tells the full-study searcher to concentrate on non-patent and foreign sources.
Briefing the search
| Supply | Effect |
|---|---|
| The patent number and the claims to target | Focus |
| The correct priority date, with the chain | Prevents a wasted search |
| Which claim elements are hardest to find | Directs effort |
| Field terminology and synonyms | Crosses the vocabulary gap |
| Art already considered during prosecution | Avoids repetition |
| Jurisdictions and languages | Scope |
| Whether the art must be documentary | Critical if a PTAB route is planned |
Say whether the art must be a patent or printed publication. Inter partes review and ex parte reexamination accept nothing else, so evidence of prior public use is worthless for those routes however strong it is.
Share the prosecution history. Art the examiner already applied is worth less, because a PTAB petition relying on it risks discretionary denial. See PTAB discretionary denial.
Search strategies that work
| Strategy | Why |
|---|---|
| Start from the claim elements | The target is known — use it |
| Search each element separately, then combine | Finds partial disclosures for §103 |
| Follow the citations backwards | The patent's own references lead elsewhere |
| Search the inventors' earlier work | Their own prior publications count |
| Search the assignee's earlier products | Their own prior art |
| Look at what the examiner cited but did not apply | May teach more than assumed |
Searching the inventors' own earlier publications is consistently productive and consistently forgotten. An inventor who published on the topic before filing may have created their own prior art.
The same applies to the assignee's earlier products. A company's own prior product line can anticipate its later patent.
Grading what comes back
| Grade | Meaning |
|---|---|
| Anticipatory | One reference discloses every element as claimed |
| Strong obviousness | Two references, clear motivation to combine |
| Moderate | Combination requires an argument |
| Background | Relevant context, not invalidating |
| Not prior art | Post-dates the priority date |
Anticipation is worth far more than obviousness. A single reference disclosing every element is a clean argument; a combination requires establishing a reason to combine and a reasonable expectation of success.
Map each candidate element by element before deciding it is strong. A reference that looks devastating in the abstract frequently lacks one claim element on close reading. See product mapping patent infringement.
Worked example: a search that changed the route
A company sued on a patent with a 2016 filing date.
| Step | Finding |
|---|---|
| Priority chain checked | Continuation — true priority date is 2012 |
| Knockout search, patents only | Nothing decisive |
| Full study commissioned | $11,000, four weeks |
| US patents | 3 candidates, all considered during prosecution |
| Foreign patents | 1 candidate, moderate |
| Conference proceedings | 1 paper from 2011, discloses 5 of 6 elements |
| Trade publication | 1 article from 2010, discloses the 6th |
What it produced
| Anticipation | No — no single reference has all six |
| Obviousness | Strong — two references, both in the same field, clear motivation |
| Documentary? | Yes — both are printed publications |
| PTAB route | Available |
| Previously considered? | No — neither was in the file |
The four-year priority correction was the critical finding. Searching from 2016 would have excluded both references entirely.
Neither reference was a patent. A patents-only search at any budget would have found nothing.
Being unconsidered mattered as much as being strong. Fresh art avoids the previously-considered ground for discretionary denial.
Using the results
| Route | What the art must be |
|---|---|
| Inter partes review | Patents and printed publications |
| Ex parte reexamination | Patents and printed publications |
| Post-grant review | Anything, within 9 months of issue |
| District court defence | Anything |
| Negotiation leverage | Anything credible |
Strong art that is not documentary still has value in negotiation even where it cannot be used at the Board.
Sort the results by route availability, not just by strength. A devastating prior-use case is worth nothing in an IPR and a great deal in court.
When the search finds nothing
A thorough search returning nothing is a real result.
| Implication | Response |
|---|---|
| Invalidity is a weak defence | Shift effort |
| Non-infringement | Map element by element |
| Design-around | Price it |
| Negotiation | Position accordingly |
| Do not file a weak petition | Failure leaves the patent stronger |
Filing a weak PTAB petition is worse than not filing. A patent that survives comes out stronger, your art is disclosed, and estoppel may attach.
The search fee is not wasted in that case. Knowing invalidity will not work redirects the budget to routes that might. See patent dispute.
Cost against the alternative
| Cost | |
|---|---|
| Knockout search | Low |
| Full invalidity study | $5,000–$20,000 |
| PTAB petition and proceeding | Substantially more |
| District court validity fight | Part of a seven-figure case |
| Settling a claim you could have defeated | Whatever was demanded |
The search is the cheapest step in the sequence and the one that determines whether the rest is worth taking.
It is also the cheapest step at renewal. An owner deciding whether to pay an $8,280 third maintenance fee benefits from knowing whether the patent would survive a challenge. See patent search cost.
Privilege
| Commissioned | Protection |
|---|---|
| Directly from a search firm | May not attract full protection |
| Through counsel | Attorney work product |
| By an in-house lawyer | Depends on role and purpose |
Route it through counsel where the results could be unfavourable to you — particularly when assessing a patent you own or are considering buying.
For a defendant searching an opponent's patent the concern is lower, since unfavourable results simply mean the defence is weak rather than exposing your own asset.
Timing against deadlines
| Deadline | Effect on the search |
|---|---|
| One-year IPR bar | Search must complete well before it |
| Post-grant review window | Nine months from issue — very tight |
| Invalidity contentions in litigation | Local rules set the date |
| Late-found art | Frequently excluded in litigation |
Late references are often excluded. Local patent rules generally require good cause to add prior art after contentions, which makes the search deadline earlier than people expect.
Work backwards from the contention date, not from the trial date. That is the point at which your art must be identified.
Patent invalidity search: the checklist
- Establish the true priority date first, following the whole continuation chain. Getting this wrong invalidates the search.
- Check whether the priority claim is supported. Unsupported priority can open a window of art.
- Run a knockout search first. It occasionally ends the question cheaply.
- Specify whether the art must be documentary, since PTAB routes accept nothing else.
- Commission non-patent literature coverage. That is where the best references are.
- Include foreign and foreign-language sources.
- Share the prosecution history so the searcher avoids art already considered.
- Map every candidate element by element before treating it as strong.
- Route the search through counsel where unfavourable results would expose your own position.
- Treat a nil result as information, and redirect the budget to non-infringement, design-around or negotiation.