For more than a decade, the answer to “we’ve been sued for patent infringement” was reflexive: file an inter partes review. The IPR was the break-glass tool of the post-AIA era — expensive, adversarial, but predictable enough to manufacture settlement leverage and, with luck, a litigation stay. That reflex has become expensive to maintain, and the numbers are starting to show it. As of 2026, the data shows challengers walking away from the IPR in numbers no one anticipated, and walking toward a procedure most of them had mothballed years ago: ex parte reexamination.
This is a structural reordering of how patent validity gets challenged in the United States, and it changes the job description for anyone responsible for clearing prior art before a fight begins.
A procedure left for dead just overtook the IPR
Consider the headline numbers. Ex parte reexamination requests hit an all-time high of 726 filings in 2025, a 66% jump over the 437 filed in 2024. Over that same stretch, IPR petition volume fell, and by early 2026 the collapse turned vertical., one four-week window in spring 2026 saw just eleven IPR petitions filed, the lowest stretch since the program launched in September 2012. For the first time in the AIA era, reexamination overtook the IPR as the leading first-stage tool for knocking out a patent.
A tool the patent bar had written off as slow, clumsy, and toothless is now the busier of the two. Understanding why is what every IP team needs to work through before its next validity challenge.
What’s actually driving the migration
The shift traces back to a USPTO that, over roughly twelve months, made the IPR dramatically harder to win before you even get to the merits, and it happened fast.
Starting under Acting Director Coke Morgan Stewart and accelerating sharply after Director John Squires was confirmed in September 2025, the agency reworked how IPR institution decisions get made. Institution authority was pulled back from three-judge panels to the Director’s office, and a March 2025 workload-management memo formalized a bifurcated process in which discretionary considerations are weighed before the merits are ever reached. The effect on the numbers is stark: PTAB institution rates dropped from roughly 68% in 2024 to under 40% by December 2025. Discretionary denials hit a record high of 607 in 2025, rising more than 630% year over year by one count. A petitioner can now spend north of $100,000 building a petition and receive a summary denial with no substantive explanation, no roadmap, no second chance on the merits.
“Settled expectations” became a trap. Patent owners can now file a briefing dedicated to why a petition should be discretionarily denied, including arguments that the challenger waited too long to act. The clearest cautionary tale is iRhythm Technologies v. Welch Allyn, where the IPR was denied in part because the petitioner had been aware of the application that matured into the challenged patent for twelve years before petitioning. The longer a patent has been in force, the more vulnerable an IPR petition is to denial on this ground alone.
The prior-art rules tightened. On July 31, 2025, the USPTO issued guidance strictly enforcing the rule that an IPR petition must supply every missing claim limitation with actual prior-art patents or printed publications. The old workaround, pointing to the patent owner’s own specification as Applicant Admitted Prior Art to fill gaps, is gone. You can no longer assert that a specification concedes a server was known; you have to find a publication that says so. For system and software patents, that single change can sink an otherwise solid challenge.
And the IPR’s estoppel still bites in a way reexamination’s doesn’t. Lose an IPR on the merits and 35 U.S.C. § 315(e) bars you from raising in court any ground you reasonably could have raised. Reexamination carries no equivalent estoppel against the requester. There is no one-year service deadline, requests can be filed anonymously through an agent, USPTO fees are a fraction of an IPR’s cost, and the institution threshold —, a “substantial new question of patentability”—, is cleared in well over 90% of well-prepared requests.
When you lay those factors against each other, the calculus shifts. The IPR has become the high-risk, high-cost option that may never reach the merits. Reexamination is the door that stays reliably open.
Where IP.com stands on this
The honest read is that this is a rational migration, not a panic, and challengers who treat it as a simple swap of one form for another will be disappointed. The shift to reexamination doesn’t lower the bar on prior art; it raises it.
In an IPR, a petitioner could lean on expert declarations, “common sense,” and applicant admissions to bridge the gaps in the art. Reexamination offers none of that latitude, and the 2025 AAPA crackdown closed the same shortcut for IPRs. Both paths now live or die on one thing: whether you have actual patents and printed publications that read on every limitation. The proceeding got easier to enter; the evidentiary burden got harder to meet. Whoever owns the most rigorous prior-art search wins, and whoever doesn’t burns a five-figure filing fee to lose slower.
That makes this a search-and-evidence problem first and a procedural problem second. It’s exactly the wrong moment to be running keyword-and-Boolean searches against patent literature alone.
Proof that the strategy works, and why prior art is the whole game
Two recent cases make the point concretely.
The first is In re Gesture Technology Partners (Fed. Cir., Dec. 1, 2025). Samsung had requested ex parte reexamination of a Gesture-owned patent covering camera-based motion sensing, while Unified Patents ran IPRs against the same claims. After the IPRs concluded and were affirmed, Gesture tried to terminate the reexamination, arguing IPR estoppel should kill it. The Federal Circuit refused, holding that § 315(e)(1) estoppel does not reach an ongoing reexamination because the requester does not “maintain” it, the Patent Office does. The practical lesson: reexamination can keep pressure on a patent even after an IPR is over, a second bite that the IPR’s own estoppel would otherwise forbid. But notice what made it work, the unpatentability finding rested on prior art that actually anticipated the claims.
The second is the broader pattern of discretionary denials like iRhythm. Those denials never reach the prior art at all; they turn the petitioner away at the threshold. Reexamination sidesteps that gauntlet entirely, but only if the request raises a genuine substantial new question, again, a function of the quality of the art you bring.
Both cases point to the same conclusion: Whether you’re entering through reexamination’s open door or trying to survive the IPR’s narrowing one, the deciding factor is no longer procedural cleverness. It’s the strength and completeness of the prior art on the page.
And lest anyone think this is a fringe academic trend, in March 2025 testimony to Congress on the reduced availability of AIA trials, Director Squires offered a four-word summary of the new regime: “there’s always reexamination.” When the head of the agency points challengers toward a specific tool, the migration is policy, not accident.
Stay ahead by making prior art your competitive edge
If the whole contest now hinges on the art, the response is to industrialize how you find it. That means going beyond patent databases and keyword strings to surface the non-patent literature, technical journals, conference proceedings, standards documents, product manuals, that so often contains the printed publication a system-type challenge requires.
This is where IP.com fits the problem directly. InnovationQ+ runs on the proprietary Semantic Gist® engine, which lets searchers describe an invention in plain language and surface conceptually relevant references across both patent and non-patent literature, including deep IEEE full-text coverage, rather than relying on brittle Boolean queries that miss art worded differently from the claim. For a challenger assembling a reexamination request or stress-testing an IPR petition against the new printed-publication-only standard, that breadth is the difference between a request that raises a substantial new question and one that doesn’t.
The flip side is just as important, and it’s where InnovationQ earns its keep. InnovationQ is IP.com’s prior-art database built specifically for defensive publishing, a single platform where innovators can both search the art and establish it. With reexamination requests climbing, the smartest patent owners aren’t just defending claims after the fact; they’re getting ahead of the challenge by publishing their own technical disclosures into a searchable, time-stamped record that becomes citable prior art the moment it goes live. That does two things at once. It blocks competitors from patenting around your work, and it builds a documented prior-art moat that makes your own portfolio harder to invalidate when someone comes looking. In a regime where every validity fight now turns on findable printed publications, a defensive publication in InnovationQ is one of the cheapest, fastest ways to put a printed publication exactly where you want it, on the record, on your terms, and working for you instead of against you.
Why this matters for innovators, R&D, and IP professionals
For in-house IP teams and outside counsel, the takeaway is procedural discipline: stop defaulting to the IPR. Pressure-test every contemplated petition against the new reality, Is a trial date too close? Is the patent old enough to invite a “settled expectations” denial? Do you need admitted prior art to make the case stick? If any answer raises a flag, reexamination is likely the smarter vehicle, and the prior-art search needs to start earlier and dig deeper than it used to.
For R&D leaders and innovators, the message runs upstream. A world where validity challenges turn entirely on findable printed publications is a world that rewards strong novelty discipline at the point of invention, and rewards defensive publishing as a low-cost way to stake out prior art against competitors.
For tech transfer offices and patent novelty consultants, this is a service opportunity. Clients facing assertion need a clear-eyed read on which post-grant vehicle actually fits their facts, and clients building portfolios need confidence that their claims can withstand the heightened evidentiary scrutiny now applied on both paths. In each case, the value you deliver traces back to the quality of the prior-art intelligence behind the advice.
The procedure changed. The deeper lesson is that patent challenges have become, more than ever, a contest of evidence, and the team with the better prior art holds the advantage.
