The Collapse of IPR—and Here’s What Founders Need to Know
Authored by Babak Akhlaghi on May 29, 2026. Between October 2024 and February 2026, the IPR institution rate collapsed from 65% to 37%.
That’s a 28% reduction in institutions. Not a gradual decline. A systemic shock.
Challengers didn’t disappear. They pivoted. Ex parte reexamination requests surged, hitting an all-time quarterly high of 198 filings in Q3 2025 alone
Then in April 2026, the USPTO responded with emergency procedural changes that fundamentally altered how patent challenges work.
Most founders don’t realize these shifts directly affect whether their patents survive challenges. This isn’t abstract patent policy. This is the difference between a patent you can enforce and a patent that gets invalidated before you ever see a courtroom.
Here’s the causal chain, what it means for your portfolio, and where patent law is heading.
The IPR Institution Rate Didn’t Decline—It Collapsed
In 2025, IPR institution decisions shifted dramatically, with the USPTO increasingly issuing discretionary denials and, in many months after June 2025, denying more petitions than it instituted.
No reasoning beyond stating “institution of inter partes review is denied.”
This stands in stark contrast to the PTAB’s historic institution rate of approximately 67% under Director Kathi Vidal between 2022 and 2024.
The shift started earlier. In March 2025, Acting USPTO Director Coke Morgan Stewart issued a memorandum creating a novel bifurcated institution process.
Step 1: The Director first determines whether discretionary denial is appropriate based on expanded Fintiv factors and PTAB staffing considerations.
Step 2: Only if discretionary denial is deemed inappropriate does the case proceed to a three-member PTAB panel to assess merits.
This two-stage gatekeeping mechanism was implemented as a workload management measure, against a backdrop of reduced PTAB staffing, including reported declines of approximately 13% in administrative patent judges and 17% in support staff.
The result was significant. Discretionary denials increased sharply under the interim framework.
By Q4 2025, the institution rate against non-practicing entities fell to 33.6%, down from 64.6% in Q1 2025. For operating companies, the rate decreased to 53.2%, down from 81.6%.
Why Challengers Flooded to Ex Parte Reexamination
When the USPTO effectively closed IPR challenges, challengers didn’t stop. They repackaged their arguments and shifted to ex parte reexamination.
EPR requests are granted at an average rate of 95% between 2020 and 2024, compared with 63% for IPR petitions.
Unlike IPRs, ex parte reexamination is not subject to Fintiv discretionary denial based on trial date proximity or parallel litigation.
For platform companies managing massive portfolios, this made EPR the one door that remains reliably open in a centralized, discretionary regime.
By the end of September 2025, the industry had filed 425 EPR requests, nearly matching the full-year total for 2024. In Q1 FY2026 alone, the USPTO recorded 223 EPR filings, an annualized rate of nearly 900, compared to 407 filings in FY2024.
Meanwhile, IPR filings plummeted from an average of 110 per month to just 66 in November 2025.
The Amazon v. Croga Case That Exposed the Loophole
Amazon filed an IPR petition against Croga Innovations. The USPTO denied the petition, noting the patent had already been subject to four reexamination requests and it was not appropriate to use office resources to challenge the patent again.
Amazon repackaged that denied IPR petition and filed it as an EPR challenge. It was instituted with the same prior art rejection.
This wasn’t an isolated tactic. Challengers realized they could take rejected IPR arguments, submit them through the EPR process, and bypass the discretionary denials that were blocking IPR institutions.
The USPTO saw the loophole and moved to close it.
The April 2026 Emergency Response: The Pre-Order Procedure
On April 1, 2026, Director Squires issued an Official Gazette Notice invoking the “extraordinary situation” provision to waive existing regulations.
The new rule allows patent owners to file 30-page pre-order papers within 30 days of being served with an EPR, before the USPTO determines whether the request raises a substantial new question of patentability.
This is unprecedented. Patent owners can now oppose EPR requests before the Office decides whether there’s an SNQ, with requesters limited to just 10-page responses.
The Notice cited “recent increased volume of ex parte reexamination requests” as justification. CRU staffing has decreased significantly over the past year, making it difficult to meet the statutory three-month SNQ determination deadline.
Under the new pre-order response, the patent owner can argue the art is cumulative if it was subject to an earlier petition that was rejected. The patent owner gets the last word to defend the patent and explain why the challenge should not be instituted.
What This Means for Founders: Your Homework Starts Today
I tell founders the same thing I would tell my startup clients who have raised over nine figures.
The days of serial challenges via IPR no longer exist. Your patent is worth something. You can take it to the bank.
But all the procedural changes in the world won’t save a patent that doesn’t provide strategic advantage.
If your patent doesn’t cover the core aspect of your technology, the features competitors care about when it comes to enforcing it, all the procedural changes won’t provide you with the leverage you need.
Here’s the evaluation framework I use with founders who want to know if their existing portfolio benefits from this shift:
First: Do your patents actually claim the core innovation that defines your competitive advantage, or are they peripheral features?
Second: Are the claims broad enough to cover competitor workarounds, or are they so narrow that design-arounds are trivial?
Third: Does your portfolio protect your current product roadmap and business model, or just technology you’ve already moved past?
Fourth: Can you afford to enforce these patents if challenged? Stronger procedural protections only matter if you’re willing to litigate.
If your patents pass these tests, the new rules are a huge win. If there are gaps in coverage or enforceability, procedural changes won’t save a fundamentally weak portfolio.
The 30-Day Window Is Non-Extendable: What Actually Works
When a patent owner gets that 30-day notice to respond before the USPTO decides whether to order ex parte reexamination, you can’t buy time. The deadline is non-extendable.
Your homework starts today, not when your patent is challenged.
Secure a capable patent counsel now. If you have serious patents and you’re looking to enforce them, a challenge will be coming. Have a strong patent firm on your side ready to jump in and defend you. Ideally, go with the patent law firm who prosecuted your patent application. They know the history better than anybody else.
Line up your expert witnesses now. Expert declarations don’t count against your 30-page limit, so use them. You don’t want the examiner faced with a one-sided expert report. When an expert with a PhD in the field says your invention is obvious to the examiner, you better have an equally qualified expert to rebut that assertion.
Put your best foot forward at this stage. Spare no expense. You want to prevent the challenge from being instituted. Otherwise, you’ll be in for a likely long battle.
The cost difference is significant. Once reexamination is ordered, it becomes a ping-pong match. You respond to one rejection, overcome it, and the examiner issues another rejection over another reference.
Budget approximately $40,000 to $60,000 for a light amendment strategy with limited Office Actions. The middle range runs $60,000 to $90,000 for typical CRU prosecution with amendments and argument. The higher end reaches $90,000 to $120,000 or more for multiple Office Actions, complex amendments, and appeal-adjacent briefing.
If you can get the response filed within a $20,000 to $40,000 budget and prevent SNQ determination, you save money and significant time.
What Actually Moves the Needle in Your 30-Page Response
Your response is limited to why the asserted references don’t raise a substantial new question of patentability.
Focus on why the cited references are cumulative over the art that has already been considered and is part of the record. Explain why the combination teaches away from the invention. Place significant focus on why the cited references alone or in combination fail to teach various features of your invention.
Put your strongest argument first, then move forward with other more nuanced distinctions.
What doesn’t matter at this stage: whether the USPTO should use its discretionary powers to reject the petition. That determination is made later if the challenge moves forward.
Don’t argue detailed motivation-to-combine analysis, teaching away beyond what is needed to show no SNQ, over-engineered claim construction fights, full 103 framework analysis, or secondary considerations. Reserve these for later if the SNQ determination is made against you.
If SNQ is ordered, interview with the examiner. Interview, interview, interview. It helps break barriers. Don’t argue. You don’t win by arguments. If you prove the examiner is wrong in his position, that may cause resentment. The examiner may be convinced, but you haven’t won him to your way of thinking. You may have just alienated him further.
Instead, understand the examiner’s point of view. Ask, “Can you please help me understand the rejection better?” The examiner wants to work with you. They want to approve patentable inventions. Explain the invention to the examiner and ask them to understand how the prior art teaches the invention.
The Bigger Picture: We’re Entering an Era of Stronger Patent Rights
In 2024 alone, U.S. patent juries awarded over $4.3 billion in damages, with average jury verdicts at record highs.
Non-practicing entity lawsuits increased roughly 30% in early 2025.
Recent patent litigation has produced increasingly large jury awards, with many software and technology cases falling in the tens of millions and frequently reaching into the nine figures. In September 2025, Anonymous Media Research won a $78.5 million verdict against Samsung Electronics in the Eastern District of Texas over media usage and automatic content recognition in smart TVs.
The signal continues to be the same from the USPTO: once we issue a patent, we’re going to stand behind it. It’s not so easy to invalidate it.
When patents become harder to kill, investors are more willing to invest in them or in the companies behind them, knowing their investment is safe. Their investment won’t be easily knocked out by bigger companies when it comes to enforcement.
Seasoned investors and bigger companies with strong IP representation already know this. They’re paying attention. Average patent awards were significantly up in 2025 compared to 2024.
A number of founders I deal with come to us because investors told them to get a patent and come back.
The same scenario plays out in licensing negotiations with big companies. They can no longer take the posture of “sue me, I’m going to invalidate your patents with IPR/EPR.” The days of serial challenges to patents to oblivion are gone. Now they have to entertain licensing negotiations or risk being exposed to expensive litigation.
What I’m Watching Next: AI in Examination and Patent Eligibility Guidance
In the next two to three years, I don’t anticipate the pendulum swinging back.
With the number of challenges to patents decreasing as a result of their lack of effectiveness, we’ll see more resources directed to cases on appeal. The backlog of patent applications under examination pending appeal will decrease significantly, a trend already emerging.
I believe the next shift will be the use of AI in examination of cases. Starting next few months, we’re going to see AI tools that help examiners provide better examination of cases both on patentability and patent eligibility grounds.
I believe these AI tools will be better able to review case law and USPTO guidance and guide examiners in issuing or not issuing patent eligibility rejections.
We’re also going to see further guidance on patent eligibility rejections, which are still being issued at the USPTO at an alarming rate with no consistency despite recent guidance.
The patent system used not to be political. With the introduction of these post-issuance challenges, it has become political. Whoever heads the office has enormous powers on the enforceability or lack thereof of patents.
Once a new administration comes in, depending on who replaces the existing directors, we may go back to an era more favorable to patent challengers, big corporations.
But for now, the door is closing on serial challenges. The procedural tweaks directly attack the heart of challengers’ ammunition. Take those tools away, and they’re left with what has been the tradition: challenging issued patents in courts.
And at the district court, it’s much harder to invalidate a patent than at the USPTO.
The Bottom Line for Founders
I once represented a patent owner who had nothing but issued patents. The patents were strategically developed. When licensing negotiations failed with big companies, the big companies challenged the patents under EPR. SNQ was instituted.
You could tell all of the client’s money was tied up with these patents. They didn’t have enough to pay our legal bills at the time, but we were able to secure a notice of allowance for them.
As soon as the notice of allowance was issued, the licensing deals started coming. They happily paid our legal bills, which were significant.
It was wonderful to see their innovation and hard work pay off.
Some argue that NPEs are bad. I ask why. They own a property right, a patent. Why can’t they enforce it? It’s like saying a landlord who intends to rent the house he buys or the land he develops is bad. There’s nothing bad about it. That’s how property rights work.
The value of patents just increased. If you want to tell a founder who says “Big Tech will just IPR my patents anyway, so why bother?” that advice is outdated, tell them this:
That was true before 2025, but the landscape has completely shifted under new leadership at the USPTO.
If you want to challenge a patent with an IPR now, you have to show the patent has not been subject to another challenge before. You have to show who is the entity challenging the patent. No more hiding behind the curtain. You have to show every element of the invention in the cited references. No more hand-waving or expert testimony to fill in the gaps.
Even if you can show all of these, the USPTO still has discretionary powers to deny the challenge, and they’ve been using it increasingly more, particularly when the patent has been issued for a few years and has gained settled expectation.
If you’re a founder thinking about enforcing your portfolio, look at those patents that have been in existence longer than the newly issued ones. Focus on patents issued more than 5-6 years ago.
Your patent, the one you fought so hard for, the one you paid so much for, is actually worth something.
You can take it to the bank.
For a detailed breakdown of how to use the new 30-day response window, see our guide here → [link]
