Checklist · Intellectual Property

Filing an Inter Partes Review Petition: A Practical Checklist

35 U.S.C. § 311(c)

Type: Checklist
Jurisdiction: National
Date: 2024-09-27
Topic: Intellectual Property

Inter partes review is the most powerful validity tool a patent defendant has, and it is also the easiest to forfeit. This checklist walks a petitioner through every gate: confirming the patent is IPR-eligible, calendaring the Section 315(b) one-year bar from service of an infringement complaint, naming every real party in interest and privy, finding prior art that qualifies as patents or printed publications, structuring non-redundant Section 102 and 103 grounds, building limitation-by-limitation claim charts, procuring an expert declaration that will survive Section 42.65, drafting to the reasonable-likelihood standard inside a 14,000-word cap, confronting Fintiv and the parallel district-court trial date, paying the fees, and serving the patent owner. It closes with the estoppel you are buying under Section 315(e) and the stay motion you should be filing in the district court. Every phase carries a "why it matters, and the traps" note drawn from the ways real petitions die. Use it as a filing-desk instrument, not a treatise.

Filing an Inter Partes Review Petition: A Practical Checklist

Intellectual Property > Patent Litigation | Published 27 September 2024 - Updated 6 February 2026 | UpLaw - uplaw.us

An inter partes review petition is the rarest thing in American litigation: a chance to have a specialized tribunal cancel your opponent's patent, on a fixed clock, on a preponderance standard, for the price of a mid-sized discovery dispute. It is also a one-year offer. The statute starts the clock the day the complaint is served, and it does not care that you were still negotiating.

Keywords: inter partes review, IPR petition, PTAB, one-year bar, real party in interest, privy, printed publication, claim charts, expert declaration, reasonable likelihood, Fintiv, discretionary denial, IPR estoppel, motion to stay, claim construction

How to use this checklist

An IPR petition is not a brief you write; it is a filing you assemble. Fourteen thousand words, a declaration, an exhibit list, a fee, mandatory notices, proof of service — all of it complete on the day it goes in, because unlike a district-court complaint, an IPR petition cannot be amended into shape. What you file is what the Board reads.

The phases below track the order in which the work actually has to happen. Two of them — Phase 2 (the one-year bar) and Phase 3 (real parties in interest) — are jurisdictional in effect: get them wrong and nothing else you did matters. Do those first, on day one, before you have read a single piece of prior art.

Each phase gives you check items, then a "why it matters, and the traps" note drawn from the ways real petitions actually die. If you want the strategic view of how IPRs function inside a larger campaign — as leverage, as a settlement lever, as one front among several — Landmark Patent Wars: Strategic Lessons From High-Profile Case Studies is worth an hour. If you are new to the vocabulary, Patents for Inventors and Founders: Frequently Asked Questions covers the basics this checklist assumes.

For the doctrine underneath this procedure, read our companion guide, Inter Partes Review and PTAB Practice: A Comprehensive Guide. For where IPR sits in the larger defense of a patent case, see Litigating Patent Infringement: A Roadmap and Research Guide and The Utility Patent Toolkit: From Invention to Enforcement. If the patent system itself is unfamiliar ground, start with The U.S. Patent System: An Overview and How U.S. Patents Work: A Comprehensive Practical Guide.

Our running example. Kestrel Fabrication, Inc. builds automated warehouse sorting equipment. On 3 March 2025 it was served with a complaint by Anvil Ridge Licensing LLC, a non-practicing entity asserting the fictional "'842 patent" on a conveyor-diverter control scheme. Kestrel's litigation counsel spent seven months on venue motions and a failed mediation. In October 2025 Kestrel finally engaged IPR counsel, who calendared the bar date as 3 March 2026, engaged an expert, and began the prior-art search. Then, on 6 January 2026, the real-party-in-interest review surfaced something nobody had looked for — and the deadline moved seven weeks closer, to a date eight days away. Kestrel filed on 9 January 2026. It survived only because the work had already been done. We follow Kestrel through every phase below; the near-miss is the point of the story.

Phase 1 — Confirm IPR eligibility: the patent, its filing date, and the nine-month PGR window

[ ] Confirm the challenged patent has issued (IPR reaches issued patents only, not applications).
[ ] Pull the face of the patent and determine the effective filing date of every challenged claim.
[ ] Classify the patent as pre-AIA (earliest effective filing date before 16 March 2013) or post-AIA (on or after that date).
[ ] If the patent is post-AIA, confirm that either (a) nine months have elapsed since issuance, or (b) a post-grant review was instituted and has terminated. 35 U.S.C. § 311(c).
[ ] If the patent is pre-AIA, note that no PGR window applies — IPR is available from the day of issuance.
[ ] Confirm your grounds are limited to § 102 and § 103, on patents and printed publications only. 35 U.S.C. § 311(b).
[ ] Confirm you have not previously filed a civil action challenging the validity of a claim of the patent. 35 U.S.C. § 315(a)(1). (A counterclaim of invalidity does not count. § 315(a)(3).)
[ ] Check for a prior IPR by you, an RPI, or a privy on the same patent — a follow-on petition invites discretionary denial under the General Plastic factors.
[ ] Check the prosecution history: if your best art was already before the examiner, expect a § 325(d) challenge under the Advanced Bionics framework, and plan to explain what the examiner got wrong.

Why it matters, and the traps. IPR is a narrow instrument. It cannot reach § 101 eligibility, § 112 indefiniteness or written description, on-sale or public-use art, or prior-art systems and products. If your best invalidity theory is that the claims are abstract under Alice, IPR is the wrong forum — that fight belongs in the district court on a Rule 12 motion (see Section 101 Eligibility After Alice: A Practical Checklist). If your best art is a competitor's product that was on sale in 2009 but never described in any publication, IPR cannot hear it — and, usefully, IPR estoppel will not take it away from you either.

The nine-month trap is real and asymmetric. A patent that issued last month from a post-AIA family cannot be attacked by IPR yet. Petitioners sometimes discover this after building the whole petition. Calendar the issue date, add nine months, and confirm before you spend a dollar on prior art.

The § 325(d) trap is subtler. The Board will not gladly re-run a rejection the examiner already considered and withdrew. If your primary reference is on the face of the patent, you must show material error by the Office — a limitation the examiner overlooked, a teaching the applicant mischaracterized, a combination never actually considered. The prosecution history is therefore Phase 1 reading, not Phase 6 reading; our Prosecuting a Patent Application: A Roadmap and Research Guide explains what the file wrapper will show you, and Responding to a Patent Office Action: A Practical Checklist shows you the moves the applicant was making when it wrote the statements you are about to quote back.

One more thing to look for while you are in the file wrapper: if the applicant withheld material art or misrepresented a reference to get the claims allowed, you may be holding an inequitable conduct defense — which the Board cannot hear, but the district court can, and which renders the whole patent unenforceable if proven. IPR estoppel will not touch it. See Inequitable Conduct and the Duty of Candor: Patent Law's Atomic Bomb and Proving Inequitable Conduct in Patent Prosecution: A Comprehensive Guide. Read the file wrapper once, for both purposes.

Phase 2 — CALENDAR THE ONE-YEAR BAR under Section 315(b)

[ ] Identify the exact date of service of the complaint alleging infringement of the patent — service, not filing, not receipt of a demand letter.
[ ] Calendar the § 315(b) bar date: one year from service, on the petitioner, on any real party in interest, and on any privy of the petitioner.
[ ] Repeat that calculation for every entity in your corporate family that has been served, and for every customer, supplier, or indemnitee whose interests you control or fund.
[ ] Treat a complaint that was voluntarily dismissed without prejudice as still starting the clock. Click-to-Call Technologies, LP v. Ingenio, Inc., 899 F.3d 1321 (Fed. Cir. 2018).
[ ] Back-schedule from the bar date: prior-art search (weeks), expert engagement and declaration (4–8 weeks), drafting and claim charts (4–6 weeks), internal review and filing mechanics (1–2 weeks).
[ ] Set an internal filing target at least 14 days before the statutory bar. Filing systems fail; the statute does not forgive.
[ ] If the clock has already run, evaluate joinder under § 315(c) to a timely-instituted IPR filed by someone else — the only recognized escape hatch (§ 315(b)'s final sentence).
[ ] If joinder is unavailable, pivot to ex parte reexamination, which has no one-year bar, or preserve invalidity for the district court.

Why it matters, and the traps. This is the deadline that ends cases. It is not tolled by settlement talks, by a stay, by an amended complaint, or by the fact that the first complaint was dismissed the week after it was served. And under Thryv, Inc. v. Click-To-Call Technologies, LP, 590 U.S. 45 (2020), the Board's application of the time bar is not appealable — § 314(d) forecloses review. If the Board says you are barred, you are barred, full stop.

Kestrel was served on 3 March 2025, so counsel calendared 3 March 2026 and back-scheduled from a 17 February 2026 filing target. Working backward from mid-February, the twelve weeks of real work — five for the search, four for the expert declaration, three for drafting and charts — meant the project had to start no later than late November 2025. Kestrel started in October. That five-week cushion is the only reason this story has a happy ending, because the bar date Kestrel had calendared was wrong, and Phase 3 is where it found out.

Date	Event	Effect on the clock
12 January 2025	Anvil Ridge serves Wren Controls (Kestrel's indemnitor)	Starts a § 315(b) clock that Kestrel does not yet know is its own
3 March 2025	Anvil Ridge serves Kestrel	Apparent bar date: 3 March 2026
Oct. 2025	IPR counsel engaged; search and expert begin	12 weeks of work commences
6 January 2026	RPI review identifies Wren as an RPI	True bar date: 12 January 2026 — eight days out
9 January 2026	Petition filed, complete, with fees and service	Filed with three days to spare

The lesson is not that Kestrel was lucky. It is that a petitioner who begins the work on the apparent deadline has no margin to absorb the discovery that the apparent deadline was never the real one. Build the cushion in Phase 2, and you can survive a Phase 3 surprise. Skip it, and a single indemnitor ends the case.

The commonest fatal error in this phase is exactly that forgotten affiliate. A complaint served on a parent, a subsidiary, or a co-defendant you are indemnifying starts your clock if that entity is an RPI or privy. Run the corporate tree before you run the prior art. Related: Patent Infringement, Claims, and Defenses: A Practical Guide for Businesses and Counsel and, for the shape of the litigation you are trying to escape, The Patent Infringement Lawsuit From Filing to Appeal: A Comprehensive Guide.

Phase 3 — Identify every real party in interest and privy

[ ] List every entity that funded, directed, controlled, or will benefit from the petition.
[ ] Include parents and subsidiaries that exercise control over the defense, indemnitors paying the freight, and any joint defense group member steering strategy.
[ ] Consider membership organizations and defense aggregators — Applications in Internet Time, LLC v. RPX Corp., 897 F.3d 1336 (Fed. Cir. 2018), adopted a broad, flexible, fact-driven RPI inquiry.
[ ] File the mandatory notices identifying each RPI, related matters, lead and back-up counsel, and service information. 37 C.F.R. § 42.8.
[ ] Cross-check: does naming that RPI create a § 315(b) problem (because the RPI was served more than a year ago)? If so, the RPI question and the time-bar question collapse into one.
[ ] Update the notices within 21 days of any change. 37 C.F.R. § 42.8(a)(3).

Why it matters, and the traps. RPI identification is a statutory requirement of a complete petition. 35 U.S.C. § 312(a)(2). Petitioners are tempted to under-name — to leave off the indemnitor who is writing the checks — precisely because naming that party would trigger the one-year bar. The Board and the Federal Circuit have seen this move. The petitioner's identification carries a rebuttable presumption of correctness (Worlds Inc. v. Bungie, Inc., 903 F.3d 1237 (Fed. Cir. 2018)), but once the patent owner produces evidence, the burden of persuasion stays with the petitioner. An incorrect RPI list that conceals a time-barred party is the worst outcome available: the petition dies, and the concealment colors everything.

Here is Kestrel's near-death experience. Its supplier, Wren Controls, sold the diverter firmware, has a contractual duty to indemnify, is funding half of Kestrel's defense, and sat in on the invalidity strategy calls. Wren is an RPI on any honest reading of Applications in Internet Time. And Wren was served with its own complaint on 12 January 2025 — seven weeks before Kestrel. Because § 315(b) runs from service on the petitioner or any real party in interest, Kestrel's true bar date was 12 January 2026, not 3 March 2026.

Kestrel's counsel found this on 6 January 2026, eight days out, while assembling the § 42.8 mandatory notices. Two facts saved the petition. First, the substantive work was nearly finished, because the search and the expert had started in October. Second, counsel did not try to argue Wren out of RPI status to buy back the seven weeks — the tempting move, and the fatal one. Naming Wren cost Kestrel its comfortable deadline; not naming Wren would have cost Kestrel the petition, and quite possibly its credibility on everything else in it. The petition was filed complete on 9 January 2026.

Run the corporate tree and the indemnity agreements in week one. The RPI list is not a cover-page formality; it is a second, hidden statute of limitations.

Phase 4 — Prior art search: patents and printed publications only, and proving public availability

[ ] Define the priority date you must antedate, claim by claim (continuations and CIPs can have different dates for different claims).
[ ] Search U.S. and foreign patents and published applications, then non-patent literature: standards documents, conference proceedings, theses, manuals, datasheets, IETF/IEEE publications.
[ ] For each candidate reference, capture the evidence of public accessibility: library date stamps, distribution records, conference attendance, indexing, a Wayback capture with an authenticating declaration.
[ ] Confirm the reference is a printed publication under § 102 — publicly accessible to persons of ordinary skill exercising reasonable diligence. In re Hall, 781 F.2d 897 (Fed. Cir. 1986).
[ ] Meet the Board's institution-stage burden on accessibility: the petition itself must show a reasonable likelihood the reference qualifies. Hulu, LLC v. Sound View Innovations, LLC, IPR2018-01039 (PTAB Dec. 20, 2019) (precedential).
[ ] Nail down the § 102 subsection you are relying on and whether pre-AIA or AIA § 102 governs.
[ ] Prepare an authenticating declaration (librarian, custodian, archivist) for any reference whose date is not self-proving.
[ ] Discard art you cannot date and prove, however good it reads.

Why it matters, and the traps. IPR practitioners lose more good art to proof problems than to substance. A slide deck from a 2011 trade show is devastating — and worthless unless you can show it was disseminated. Hulu is the case that turned this from a trial-stage problem into a petition-stage problem: you must front the accessibility evidence in the petition, because the Board will not let you fix it in reply.

Note the reach of the search. The classical clearance search asks what is in force; the IPR search asks what was published before a date, regardless of whether it ever issued, expired, or was abandoned. Expired patents, abandoned applications, and dead foreign filings are all fair game. Our Running a Freedom-to-Operate Patent Search: A Practical Checklist covers search technique; invert its logic for invalidity, and widen it.

Phase 5 — Select and structure the grounds; avoid redundancy

[ ] Frame each ground precisely: "Claims 1, 4, and 9 are unpatentable under § 103 over Okabe in view of Lindqvist."
[ ] Order grounds by strength, not by elegance. The Board reads the first ground with the most attention.
[ ] Limit yourself to three to five grounds. Ten thin grounds read as a petitioner who has none.
[ ] For each § 103 ground, articulate the motivation to combine and the reasonable expectation of success — do not let the expert supply them by assertion alone.
[ ] Apply the Graham factual inquiries and KSR's flexible approach: Graham v. John Deere Co., 383 U.S. 1 (1966); KSR International Co. v. Teleflex Inc., 550 U.S. 398 (2007).
[ ] Anticipate secondary considerations (commercial success, long-felt need, copying) and preempt the strongest one.
[ ] Confirm each reference's prior-art status against every challenged claim — not just claim 1.
[ ] Decide, deliberately, which claims to challenge. Under SAS Institute Inc. v. Iancu, 584 U.S. 357 (2018), institution is all-or-nothing across the claims you challenge: the Board must address every one in the final written decision.

Why it matters, and the traps. SAS changed the arithmetic of petition drafting. Before SAS, a petitioner could throw claims at the wall and let the Board institute on the good ones. Now, if the Board institutes, it institutes on all challenged claims and all grounds — and § 315(e) estoppel then attaches to every claim that reaches a final written decision. Challenging a marginal dependent claim you cannot actually invalidate does not just waste words; it buys you an estoppel on a claim you would rather have fought in district court.

Kestrel's grounds show the discipline. Anvil Ridge asserted claims 1–20 of the '842 patent, but only claims 1, 4, 9, and 12 read on Kestrel's sorter. Counsel challenged those four and no others — declining, deliberately, to challenge claims 15–20, which were weak-looking but irrelevant and would have cost excess-claims fees and bought a gratuitous § 315(e) estoppel.

Ground	Statute	Art	Claims	Why it is here
1	§ 103	Okabe in view of Lindqvist	1, 4, 9	Lead ground. Okabe teaches the diverter array; Lindqvist supplies the closed-loop timing controller and states the design need expressly.
2	§ 102	Okabe alone	1	Straight anticipation if the Board reads "control module" as broadly as Anvil Ridge does in the district court. A trap laid with the plaintiff's own construction.
3	§ 103	Okabe + Lindqvist + Sasaki	12	Sasaki supplies only the dependent limitation (the encoder feedback loop) — a targeted ground, not a shotgun.

Three grounds, four claims, one primary reference. Note that Ground 2 exists only because the patent owner's own infringement contentions demand a broad construction of "control module"; if Anvil Ridge narrows that term to escape Okabe, it narrows its way out of the infringement case. That is the construction whipsaw run in the petitioner's favor — see Phase 8.

The § 103 architecture is where petitions are won. For the analytic framework, read Obviousness Under Section 103: A Comprehensive Guide to the Analysis; to understand how the patent owner will push back, read the mirror-image checklist, Defeating a Section 103 Obviousness Rejection: A Practical Checklist, and its prosecution-side cousin, Turning Patent Rejections Into Granted Claims: A Prosecution Response Guide — the arguments an applicant used to win the claims are the arguments the patent owner will now use to save them. Write your petition as if you had already read the patent owner's response — because you can.

Phase 6 — Build the claim charts: map every limitation to the art

[ ] Break each challenged claim into numbered limitations ([1.a], [1.b], [1.c] …) and use those labels everywhere.
[ ] For each limitation, quote the specific disclosure in the reference: column and line for patents, page and paragraph for publications, with figure callouts.
[ ] Never chart a limitation to a conclusory string cite. "Okabe, Figs. 2–7, 4:12–9:40" is not a mapping; it is a hope.
[ ] For combinations, state which reference supplies which limitation and why a skilled artisan would put them together for that limitation.
[ ] Comply with 37 C.F.R. § 42.104(b)(4)–(5): specify where each element is found, with exhibit numbers and page/line citations.
[ ] Cite the expert declaration at each contested limitation — and make sure the declaration actually says it.
[ ] Chart every challenged claim, including dependents. Dependent claims are where petitions quietly fail.

Why it matters, and the traps. The Board reads claim charts the way a district judge reads a summary-judgment appendix: skeptically, and looking for the gap. The petitioner's most common self-inflicted wound is the "and it would have been obvious" bridge — a limitation that appears in no reference, papered over by an expert's ipse dixit. If the art does not disclose it, say so and explain, with evidence, why a skilled artisan would have supplied it anyway. Pretending the gap is not there is worse than admitting it.

Because claim scope drives every mapping, the charts and the construction section must be built together. See Claim Construction and the Markman Hearing: A Practical Checklist.

Phase 7 — The expert declaration: obviousness, motivation, and the level of skill

[ ] Engage the expert early enough to be independent — an expert handed a finished petition to sign is a gift to the patent owner.
[ ] Have the expert define the person of ordinary skill in the art (POSITA) as of the priority date, with education and experience specified.
[ ] Have the expert explain the state of the art, the technical problem, and the known solutions before touching the claims.
[ ] Require the expert to supply the underlying facts and data for every opinion. 37 C.F.R. § 42.65(a): testimony that does not disclose the underlying facts or data "is entitled to little or no weight."
[ ] Build the motivation to combine on record evidence — a teaching in the references, a known design need, market pressure — not on hindsight assembled from the claim.
[ ] Address reasonable expectation of success explicitly.
[ ] Confirm the declaration is signed, dated, and served as an exhibit, with the expert's CV.
[ ] Prepare the expert for cross-examination by deposition during the trial phase (37 C.F.R. § 42.51(b)(1)(ii) routine discovery).

Why it matters, and the traps. Section 42.65 is the PTAB's Daubert, and it bites harder than most petitioners expect. The Board routinely discounts declarations that parrot the petition verbatim, and patent owners routinely quote the parroting back. An expert whose declaration tracks the petition's headings word for word, with no independent technical narrative, gives the Board an easy reason to say the petitioner has not carried its burden. Write the declaration first; write the petition from it. On the cross-examination side of this coin, see Challenging an Expert Under Daubert and Rule 702: A Practical Checklist — the same weaknesses that sink an expert under Rule 702 sink one under § 42.65.

Phase 8 — Address claim construction where it matters

[ ] Apply the Phillips standard — the same one the district court uses. 37 C.F.R. § 42.100(b); Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc).
[ ] Construe only the terms that matter to your grounds. Gratuitous constructions burn words and create hostages.
[ ] Disclose any construction the district court has already adopted, and any construction the patent owner has advanced there — the Board considers prior constructions of record.
[ ] Watch for means-plus-function terms under § 112(f): identify the function and the corresponding structure, or explain the indefiniteness problem (knowing the Board cannot cancel on § 112 grounds).
[ ] Guard against the construction whipsaw: a broad construction that captures the art may also capture your product. Reconcile your IPR position with your non-infringement position, in writing, before filing.
[ ] Cite the prosecution history for disclaimer where it helps.

Why it matters, and the traps. Since the 2018 rule change aligned the PTAB with Phillips, the strategic gap between the two forums has narrowed — and the whipsaw risk has grown. Your petition's constructions are admissions the plaintiff will read aloud in the district court. A petitioner who argues in the IPR that "diverter assembly" is broad enough to read on Okabe, and argues in Delaware that the same term is narrow enough to exclude its own sorter, will be asked to explain the difference to a jury. The two positions must be consistent, or the inconsistency must be principled and articulable. See The Patent Infringement Lawsuit From Filing to Appeal: A Comprehensive Guide.

Phase 9 — Confront Fintiv: the parallel litigation and the trial date

[ ] Pull the district court's scheduling order and identify the trial date.
[ ] Compare it to the Board's projected final written decision date: institution + 12 months (§ 316(a)(11); 37 C.F.R. § 42.100(c)).
[ ] Work through the six Fintiv factors (Apple Inc. v. Fintiv, Inc., IPR2020-00019 (PTAB Mar. 20, 2020) (precedential)): (1) stay granted or likely; (2) proximity of the trial date; (3) investment already made by the court and parties; (4) overlap of issues; (5) identity of parties; (6) other circumstances, including the merits.
[ ] Consider a Sotera stipulation — a broad agreement not to pursue in district court any ground raised or that reasonably could have been raised in the IPR. Sotera Wireless, Inc. v. Masimo Corp., IPR2020-01019 (PTAB Dec. 1, 2020) (precedential in relevant part).
[ ] File early. Filing at month eleven of the one-year window is itself a Fintiv factor against you.
[ ] Address discretionary denial affirmatively in the petition, and be ready to brief it separately if the Office's current practice calls for it.
[ ] Verify the current USPTO discretionary-denial guidance before you file — this is the most volatile area in PTAB practice.

Why it matters, and the traps. Fintiv is where meritorious petitions go to die. The Board can decline to institute even on a strong petition, purely because a district court will reach the same issues first. And the Office's posture on discretionary denial has swung hard: a June 2022 memorandum from the Director sharply constrained Fintiv denials (protecting petitions with compelling merits, and those backed by a Sotera stipulation), and the Office rescinded that memorandum in February 2025, restoring the full discretionary analysis and layering on a separate briefing process for discretionary denial. Treat any statement about the current framework — including this one — as perishable, and confirm the Office's operative guidance and the PTAB Consolidated Trial Practice Guide before you file.

The two levers you actually control are timing and stipulation. Filing at month four rather than month eleven, and offering a broad Sotera stipulation, are the difference between institution and a two-page denial that never reaches your art. Kestrel had no timing lever left — Wren's clock put it at the twelfth hour of the window whether it liked it or not — so it spent everything it had on the other levers: a full Sotera stipulation, a merits showing it could fairly describe as compelling, and a motion to stay filed in the district court the same week, because factor one asks whether a stay has been granted or is likely. A petitioner who cannot win factor two must make factors one, four, and six carry the weight.

Phase 10 — Draft to the reasonable-likelihood standard within the word count

[ ] Write to § 314(a): a reasonable likelihood that you will prevail on at least one challenged claim.
[ ] Respect the 14,000-word limit for an IPR petition (37 C.F.R. § 42.24(a)(1)(i)) — words, not pages, and claim-chart text counts.
[ ] Comply with § 42.22 (content) and § 42.104 (grounds for standing; identification of the challenge; construction; how the claims are unpatentable; evidence relied on).
[ ] Include the word-count certification.
[ ] Front-load: statement of relief, the art, the grounds, then the claim-by-claim analysis.
[ ] Cut every sentence that does not carry a limitation, a citation, or a reason.
[ ] Format exhibits under § 42.63, with sequential exhibit numbers and an exhibit list.
[ ] Proofread the claim numbers, the exhibit numbers, and the column/line cites — twice.

Why it matters, and the traps. Fourteen thousand words is less than it sounds when the charts are inside the count. Petitioners bleed words on background, on gratuitous constructions, and on grounds they will not win. Then, at claim 14, the analysis thins to a sentence, and the Board notices. Budget words the way you budget deposition hours: allocate them to the claims and limitations that will be contested, and spend nothing on the rest.

Remember also that the reply brief is short (5,600 words) and cannot introduce new evidence or new theories. Everything you will ever say must be in the petition. A petition is not an opening brief; it is the whole case.

Phase 11 — Mandatory notices, fees, and service on the patent owner

[ ] File the mandatory notices under 37 C.F.R. § 42.8: RPIs, related matters, lead and back-up counsel, service information.
[ ] Pay the fees under 37 C.F.R. § 42.15 — a request fee and a post-institution fee, both due at filing, plus excess-claims fees for claims beyond twenty. Confirm the current USPTO fee schedule; the amounts are revised periodically and are substantial.
[ ] Note that excess-claims fees are another reason to challenge only the claims you can actually kill.
[ ] Serve the petition and all exhibits on the patent owner at the correspondence address of record (37 C.F.R. § 42.105), or as otherwise permitted, and file proof of service.
[ ] File through PTAB E2E / Patent Center and confirm the filing receipt.
[ ] Confirm the Board accords a filing date — a petition is not "filed" until it is complete. 37 C.F.R. § 42.106.
[ ] Enter counsel's powers of attorney and, for any non-registered lead counsel, the motion for pro hac vice admission.

Why it matters, and the traps. The § 42.106 filing-date rule is where the one-year bar and the mechanics collide. An incomplete petition — missing a fee, missing an RPI, missing proof of service — is not accorded a filing date. If completeness is achieved after the § 315(b) bar has run, the petition is time-barred, and no amount of good faith cures it. This is why the internal target is two weeks before the statutory date: the buffer exists to absorb exactly this failure.

Phase 12 — Plan for the preliminary response and institution

[ ] Expect the patent owner's preliminary response within three months of the filing-date accord. 35 U.S.C. § 313; 37 C.F.R. § 42.107.
[ ] Anticipate its three moves: discretionary denial (Fintiv, General Plastic, § 325(d)), printed-publication attacks, and substantive claim-chart gaps.
[ ] Consider whether to seek leave for a preliminary reply — available only on a showing of good cause, and rarely granted.
[ ] Calendar the institution decision: within three months of the preliminary response or the deadline for it. 35 U.S.C. § 314(b).
[ ] Plan the trial schedule: patent owner response, petitioner reply, patent owner sur-reply, depositions, oral hearing, final written decision within one year of institution (§ 316(a)(11)).
[ ] Prepare for a motion to amend under § 316(d) — and remember that the burden of persuasion on the unpatentability of substitute claims rests on the petitioner. Aqua Products, Inc. v. Matal, 872 F.3d 1290 (Fed. Cir. 2017) (en banc).
[ ] Note the burden of proof at trial: preponderance of the evidence (§ 316(e)) — not the clear-and-convincing standard of district court. That gap is the whole point of IPR.

Why it matters, and the traps. Institution is not a merits ruling, and a denial is not appealable (§ 314(d); Cuozzo Speed Technologies, LLC v. Lee, 579 U.S. 261 (2016)). Petitioners sometimes plan for two bites — file, see the preliminary response, refile better. That plan runs straight into General Plastic, which disfavors follow-on petitions filed with the benefit of the patent owner's roadmap, and into the one-year bar, which will usually have run. There is one petition. Write it as if there is one petition.

Phase 13 — Understand the estoppel you are buying under Section 315(e)

[ ] Know the trigger: estoppel attaches when the IPR results in a final written decision as to a claim. § 315(e).
[ ] Know the scope: you may not later assert, before the Office (§ 315(e)(1)) or in a civil action or ITC proceeding (§ 315(e)(2)), that the claim is invalid on any ground you raised or reasonably could have raised during the IPR.
[ ] Read "reasonably could have raised" broadly: it covers grounds a skilled searcher exercising reasonable diligence would have found. California Institute of Technology v. Broadcom Ltd., 25 F.4th 976 (Fed. Cir. 2022).
[ ] Confirm what estoppel does not reach: § 101, § 112, prior-art systems and products, on-sale and public-use art — grounds IPR could never have heard.
[ ] Preserve those non-estopped defenses in the district court explicitly, in your invalidity contentions.
[ ] Weigh the trade: a preponderance standard and a technically expert tribunal, in exchange for surrendering your patents-and-publications invalidity case if you lose.

Why it matters, and the traps. This is the decision, and it should be made by the client with clear eyes in Phase 1 — not discovered in Phase 13. If the printed-publication art is your only invalidity case, an IPR is close to free: you were going to lose that fight at trial under clear-and-convincing anyway. If you also have a devastating prior-use system defense and a serious § 101 argument, the estoppel costs you less than it looks like — because those grounds survive. But if your art is marginal, an IPR converts a maybe into a loss and forecloses the district-court version of the same argument. See What Counts as Patent Infringement: An Anatomy of the Section 271 Analysis for the offensive case you will still have to defend if the claims survive.

One more asymmetry worth knowing: you do not need Article III standing to file an IPR, but you do need it to appeal an adverse final written decision to the Federal Circuit. Consumer Watchdog v. Wisconsin Alumni Research Foundation, 753 F.3d 1258 (Fed. Cir. 2014); Phigenix, Inc. v. ImmunoGen, Inc., 845 F.3d 1168 (Fed. Cir. 2017). A petitioner with no competitive injury can knock a patent out — but cannot complain if it loses.

The settlement dimension. Estoppel only attaches on a final written decision. A petition that settles before the FWD — under § 317, which permits termination on joint request — carries no estoppel at all. This is why the IPR is so often a negotiating instrument rather than a validity instrument: institution is the moment the patent owner's settlement number moves, because institution means the Board has already found a reasonable likelihood that its claims are unpatentable. Many petitions are filed to be settled, not tried. Price that possibility into the decision, and understand what the asserted patent is actually worth to its owner before you make an offer — see What an IP Case Is Worth: A Guide to Intellectual Property Damages and Monetizing a Patent Through Licensing: From Valuation to Term Sheet, which is the plaintiff's side of the same table.

Phase 14 — Coordinate with the district court: move to stay

[ ] File a motion to stay the district-court case pending IPR, promptly after filing the petition (courts stay under their inherent power; Landis v. North American Co., 299 U.S. 248 (1936)).
[ ] Brief the three factors most courts apply: (1) whether a stay will simplify the issues; (2) the stage of the litigation (is discovery complete, is a trial date set); (3) whether a stay unduly prejudices the non-moving party or presents a clear tactical disadvantage.
[ ] Offer the Sotera stipulation in the stay motion too — it powerfully supports simplification.
[ ] Preserve the record: a stay granted (or seriously entertained) helps you on Fintiv factor one; make sure the Board sees it.
[ ] Maintain your litigation hold and discovery obligations while the stay motion is pending. See Issuing a Litigation Hold and Preserving ESI: A Practical Checklist.
[ ] Coordinate positions across both forums — one construction, one story, one expert theory.

Why it matters, and the traps. The stay motion and the Fintiv analysis are a feedback loop: the Board asks whether the court will stay; the court asks whether the Board will institute. Break the loop by moving early and by making the simplification argument concrete — with a stipulation, with a claim-by-claim overlap chart, with the Board's schedule laid against the court's. A petitioner who files an IPR and then does nothing in the district court has handed the Board factor one.

Remember that a stay pauses the case; it does not pause your obligations in it, and it does not pause the case you may still have to try if the claims survive. For the district-court mechanics running alongside the IPR, see Federal Civil Litigation: A Roadmap and Resource Guide and, for a smaller defendant's view of the same machinery, Federal Civil Litigation for Small Businesses: A Patent-Focused Field Guide. If your adversary is asserting the same patent family abroad, coordinate the IPR with the foreign proceedings before you take positions you cannot repeat — Cross-Border Patent Litigation: Strategies for Navigating International IP Disputes. And if the patent is standard-essential, the Fintiv and stay calculus sits inside a much larger FRAND negotiation: Standard-Essential Patents and FRAND Licensing in the 5G and IoT Era.

Common mistakes to avoid

Starting the search before running the corporate tree. The bar date is set by service on any RPI or privy. Find it first.
Under-naming real parties in interest to dodge the bar. It does not work, and it poisons the petition.
Assuming a dismissed complaint reset the clock. It did not. Click-to-Call.
Waiting out the one-year window. Late filing is a Fintiv factor against you and leaves no margin for a § 42.106 completeness failure.
Charting a limitation to "Figs. 2–7, generally." That is not a mapping.
Letting the expert sign the petition. Write the declaration first; make it independent, factual, and § 42.65-proof.
Challenging every claim. SAS makes that an estoppel purchase, not a hedge. Excess-claims fees make it an expensive one.
Ignoring the prosecution history. If the examiner already saw your art, § 325(d) and Advanced Bionics are waiting.
Filing art you cannot date. Hulu requires the accessibility showing in the petition.
Taking inconsistent construction positions across forums. The plaintiff will read your petition to the jury.
Skipping the stay motion. It costs little and moves Fintiv factor one.
Treating estoppel as a footnote. It is the price of the ticket. Price it in Phase 1.

The consolidated IPR petition checklist

[ ] Patent issued; post-AIA nine-month PGR window closed (or patent is pre-AIA)
[ ] Grounds limited to §§ 102/103 on patents and printed publications
[ ] No prior declaratory-judgment action by petitioner (§ 315(a))
[ ] § 315(b) bar date calendared from service on petitioner, every RPI, and every privy
[ ] Internal filing target set ≥ 14 days before the bar
[ ] All RPIs and privies identified; § 42.8 mandatory notices drafted
[ ] Prior-art search complete; public accessibility proven for every non-patent reference
[ ] Authenticating declarations obtained where dates are not self-proving
[ ] Three to five non-redundant grounds selected and ordered by strength
[ ] Motivation to combine and reasonable expectation of success supported by record evidence
[ ] Claim charts map every limitation of every challenged claim to specific disclosure
[ ] Expert declaration independent, factual, § 42.65-compliant, signed and dated
[ ] Phillips constructions limited to terms that matter; consistent with district-court positions
[ ] Fintiv addressed head-on; Sotera stipulation prepared; current USPTO discretionary-denial guidance confirmed
[ ] Petition ≤ 14,000 words, with certification; § 42.104 content complete
[ ] Exhibits numbered, exhibit list attached (§ 42.63)
[ ] Request fee + post-institution fee + excess-claims fees paid (current § 42.15 schedule)
[ ] Petition and exhibits served on the patent owner; proof of service filed (§ 42.105)
[ ] Filing date accorded (§ 42.106)
[ ] Motion to stay filed in the district court
[ ] Estoppel consequences under § 315(e) explained to the client in writing

Primary Authority

Statutes

35 U.S.C. § 311 — inter partes review; grounds limited to §§ 102/103 on patents and printed publications; timing and the nine-month PGR window
35 U.S.C. § 312 — contents of the petition; real parties in interest; particularity
35 U.S.C. § 313 — preliminary response by the patent owner
35 U.S.C. § 314 — institution: reasonable likelihood of prevailing on at least one claim; timing; no appeal
35 U.S.C. § 315 — relation to other proceedings: (a) prior civil action; (b) one-year bar; (c) joinder; (e) estoppel
35 U.S.C. § 316 — conduct of the proceeding; motions to amend; preponderance standard; one-year pendency
35 U.S.C. § 317 — settlement and termination before a final written decision (no estoppel attaches)
35 U.S.C. § 318 — final written decision and certificate
35 U.S.C. § 325(d) — discretion where the same art or arguments were previously before the Office
35 U.S.C. §§ 102, 103 — anticipation and obviousness

Regulations

37 C.F.R. § 42.8 — mandatory notices
37 C.F.R. § 42.15 — fees
37 C.F.R. § 42.22 — content of petitions and motions
37 C.F.R. § 42.24 — type-volume limits (14,000 words for an IPR petition)
37 C.F.R. §§ 42.100–42.107 — IPR-specific rules: Phillips construction (§ 42.100(b)), petition content (§ 42.104), service (§ 42.105), filing date (§ 42.106), preliminary response (§ 42.107)
37 C.F.R. §§ 42.62, 42.63, 42.65 — evidence, exhibits, and expert testimony

Cases and precedential PTAB decisions

SAS Institute Inc. v. Iancu, 584 U.S. 357 (2018) — all-or-nothing institution
Cuozzo Speed Technologies, LLC v. Lee, 579 U.S. 261 (2016) — institution decisions unreviewable
Thryv, Inc. v. Click-To-Call Technologies, LP, 590 U.S. 45 (2020) — § 315(b) time-bar rulings unappealable
Oil States Energy Services, LLC v. Greene's Energy Group, LLC, 584 U.S. 325 (2018) — IPR is constitutional
Click-to-Call Technologies, LP v. Ingenio, Inc., 899 F.3d 1321 (Fed. Cir. 2018) — dismissal without prejudice does not reset the one-year clock
Applications in Internet Time, LLC v. RPX Corp., 897 F.3d 1336 (Fed. Cir. 2018) — broad RPI inquiry
Worlds Inc. v. Bungie, Inc., 903 F.3d 1237 (Fed. Cir. 2018) — RPI burden framework
California Institute of Technology v. Broadcom Ltd., 25 F.4th 976 (Fed. Cir. 2022) — § 315(e)(2) estoppel reaches grounds that reasonably could have been raised
Aqua Products, Inc. v. Matal, 872 F.3d 1290 (Fed. Cir. 2017) (en banc) — motions to amend
Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc) — claim construction
KSR International Co. v. Teleflex Inc., 550 U.S. 398 (2007); Graham v. John Deere Co., 383 U.S. 1 (1966) — obviousness
In re Hall, 781 F.2d 897 (Fed. Cir. 1986) — public accessibility of a printed publication
Apple Inc. v. Fintiv, Inc., IPR2020-00019 (PTAB Mar. 20, 2020) (precedential) — discretionary denial factors
Sotera Wireless, Inc. v. Masimo Corp., IPR2020-01019 (PTAB Dec. 1, 2020) (precedential in relevant part) — broad stipulations
General Plastic Industrial Co. v. Canon Kabushiki Kaisha, IPR2016-01357 (PTAB Sept. 6, 2017) (precedential) — follow-on petitions
Advanced Bionics, LLC v. MED-EL Elektromedizinische Geräte GmbH, IPR2019-01469 (PTAB Feb. 13, 2020) (precedential) — § 325(d) framework
Hulu, LLC v. Sound View Innovations, LLC, IPR2018-01039 (PTAB Dec. 20, 2019) (precedential) — printed-publication showing at institution
Consumer Watchdog v. Wisconsin Alumni Research Foundation, 753 F.3d 1258 (Fed. Cir. 2014); Phigenix, Inc. v. ImmunoGen, Inc., 845 F.3d 1168 (Fed. Cir. 2017) — standing to appeal
Landis v. North American Co., 299 U.S. 248 (1936) — inherent power to stay

Secondary Sources

USPTO, PTAB Consolidated Trial Practice Guide — the operative procedural manual for PTAB trials. Check for the current edition and for any superseding Office guidance or Director memoranda on discretionary denial, which have changed repeatedly and are the least stable part of this practice.
USPTO, Manual of Patent Examining Procedure (MPEP) — for reading the file wrapper and understanding what the examiner was doing during prosecution.
Chisum on Patents — the standard treatise on patent doctrine, including anticipation, obviousness, and printed-publication status.
The USPTO fee schedule (implementing 37 C.F.R. § 42.15) — verify the current amounts before filing; they are revised periodically and are substantial.

Related UpLaw Resources

Related Guides & Articles

Related FAQ

Patents for Inventors and Founders: Frequently Asked Questions

Related Checklists

Related Toolkits

This checklist is general legal information from UpLaw (uplaw.us), not legal advice, and does not create an attorney-client relationship. PTAB practice — and especially the Office's approach to discretionary denial — changes frequently; verify the current statutes, rules, fee schedule, and USPTO guidance and consult qualified patent counsel before acting.