THE EARLIER RESTRICTIONS REMAIN
Since 2025, petitioners have faced a separate discretionary review, renewed attention to parallel litigation, and settled-expectations arguments based on patent age and the parties’ history. These considerations can defeat an IPR despite strong prior art.[2] That risk can make filing unattractive before a panel ever considers the merits.
This can produce self-selection. A company may forgo an IPR because it expects discretionary denial, particularly for an older patent, rather than because its prior art is weak. The remaining petitions are selected for their prospects of clearing both the policy screens and the merits review. A higher institution rate among that smaller group would not necessarily mean greater access. Filing data cannot show how many companies decided not to file.
As Figure 1 shows below, IPR filings fell from 1,025 in January 2025 through September 2025 to 271 during the same months of 2026, a 73.6% decline. The denial share of institution decisions rose from 54.9% to 60.7%. Institutions fell from 419 to 179, a 57.3% drop, leaving a much smaller trial pipeline.[3]
TWO HURDLES BEFORE TRIAL
A petition that survives the director’s initial review will go to a three-member PTAB panel for a separate merits-based institution decision. Passing the first review does not secure a trial. The panel must independently apply the institution standard and the memorandum’s merits-related requirements. The memorandum also requires PTAB institution decisions to briefly address the merits-based considerations the panel found important, although they need not address every ground or argument.[4]
The panel must assess the showing across a substantial majority of the claims challenged in the petition. A strong challenge to one claim cannot carry numerous weaker challenges into trial. The memorandum does not define “substantial majority” numerically.[5]
The IPR statute requires a reasonable likelihood of success on at least one challenged claim. The memorandum reinforces a discretionary consideration addressing the broader claim set.[6] Petitioners therefore need to choose claims carefully and support each challenge.
The preliminary response can also prevent institution. If the patent owner presents a reasonable argument, backed by reliable, credible evidence, that the showing fails for a substantial majority of challenged claims, the panel is directed to deny review. If reasonably supported arguments and evidence for a particular claim are evenly balanced, the panel must treat the institution showing as insufficient for that claim.[7] Close calls can therefore undermine the showing across the challenged claims.
A DIFFERENT TRIAL PANEL
None of the institution-panel judges will sit on the trial panel.[8] This separation may help address concerns about confirmation bias by ensuring that the judges conducting the trial did not make the preliminary institution determination. Petitioners must therefore persuade one panel to institute and a different panel to rule for them on the full trial record.
TURNOVER AND TRAINING
Fiscal-year-end administrative patent judge (APJ) headcounts were 234 in 2022, 225 in 2023, 229 in 2024, and 204 in 2025.[9] These totals show a net decline, but do not capture changes in the PTAB’s experience.
Senior departures included former Vice Chief Melissa Haapala and former Deputy Chief Jackie Wright Bonilla, who moved to private practice in April and September 2025, respectively.[10] Those departures took substantial PTAB experience into private practice. Headcount alone does not capture that loss.
The November memorandum describes PTAB STRONG training on the reasonable-likelihood standard and the director’s institution approach, as well as judges’ work with the director on more than 325 institution decisions.[11] Panels regaining authority have already been working within that framework. The delegation does not restore earlier practice.
COMPLETE MERITS WINS ARE LESS COMMON
The contraction in trial access coincides with a lower complete-win share among cases reaching a final written decision (FWD). For 2022 filings, 69.7% of FWDs recorded within 24 months found all adjudicated claims unpatentable. The share was 62.4% for 2023 and 61.1% for January 2024 through October 2024, an 8.6-percentage-point decline from the first group to the last (Figure 2). A smaller trial pipeline and a lower complete-win share at the FWD stage put separate pressures on IPR’s usefulness.
EXPECT LOW VOLUMES TO PERSIST
The return of panel authority leaves discretionary screening and settled-expectations arguments in place, reinforces the substantial-majority showing, and changes the trial judges. Panels will apply that framework after training on the director’s approach. We therefore expect filings and institutions to remain low or fall further. That is a forecast based on the combined requirements; the November changes have not yet produced observable outcomes.
PRACTICAL IMPLICATIONS
Petitioners should assess discretionary-denial risk before filing, select challenged claims carefully, and support each challenge. Narrowing the claim set must still serve the parallel litigation strategy.
Additionally, patent owners should address both discretionary grounds and claim-specific weaknesses, supported by credible evidence. After institution, both parties should prepare for a fresh assessment by the trial panel.
Note that research cutoff is November 11, 2026. The November policy’s effect on ultimate outcomes cannot yet be observed. The expectation of continued low filing and institution volumes is a forecast based on the combined requirements.