Sept. 2, 2026

Last September, after John Squires was confirmed as Director, I wrote that we would see whether the then-emerging line of discretionary dismissals of IPR petitions “is to be extended or will itself be tempered.” The Federal Circuit answered the reviewability half on November 6. The 2026 filing and institution numbers supply the rest of the picture for now.

Discretionary non-institution remains a central feature of AIA trial practice. Institution decisions are concentrated at the Director level, and AIA petition volume is down. This site covers AIA trials, reexamination, prosecution, and enforcement. The current mix — more Chapter 30 requests, fewer IPRs — reflects how parties are using the available tools under current institution practice. It is more a reflection of institutional discretion than a ranking of the statutes. Institution guidance can change; the cases before you have to be met with all tools available.

Two denominators
Figures from the USPTO’s PTAB Trial Statistics, June 2026 (FY26 through Q3) and Unified Patents’ first-half 2026 report tell the story:

• 395 AIA petitions FY26 YTD (355 IPR / 40 PGR), against 1,433 in FY25 and 1,288 in FY24. PGR is about 10% of the docket. It did not collapse with IPR.
• Effective institution 38% (314 of 823 decided). Not the “66%” labeled YTD rate on the same PDF (merits-weighted). FY24 was 68%; FY25 was 50%.
• Director’s discretionary considerations: 681 petitions, 407 denied (60%), 274 referred. About three in five never get a merits look.
• Merits after referral: 416 petitions, 314 instituted (75%).
• Electrical/computer: 66% of filings (261 of 395), 34% instituted (178 of 523). Chemical and bio/pharma about 51–52%.
• Second quarter 2026: 57 IPRs vs. 351 in Q2 2025. First-half 2026 PTAB petitions 201, down 72.5% from 732 in the first half of 2025. Monthly IPRs: 131 in Jan. 2025, 22 in June 2026.
• First-half 2026 mix: reexam 74.7% of post-grant filings (594), IPR 21.9% (174). First-half 2025 was the reverse. Q2 reexam requests 336, an all-time quarterly high.

The mix reflects demand under current institution practice.

November 6, then Apple
In the November 6 post, the Federal Circuit denied mandamus in Motorola (No. 2025-134), SAP (Nos. 2025-132, -133), and Google (No. 2025-144). Motorola was designated precedential. What stuck is familiar but worth stating again: § 314(a) discretion, § 314(d)’s “final and nonappealable” language, and mandamus that is “ordinarily unavailable.” Interim guidance created no property interest in a merits look, and filing-cost reliance is not due process. APA attacks aimed at vacating a particular non-institution decision are treated as an end-run around § 314(d), though a freestanding APA case remains a possible path “apart from” a specific institution decision.

Apple v. Squires, No. 2024-1864 (Feb. 13, 2026) (precedential), then held that NHK–Fintiv is effectively a policy statement, not a legislative rule, and that there is no entitlement to IPR. Non-institution leaves the parties in district court and in Chapter 30. The October 17, 2025 NPRM (RIN 0651-AD89) “has no current effect.” A final-rule package was at the Office of Information and Regulatory Affairs (OIRA) as of July 22, 2026; it is still not a final rule.

VirtaMove, No. 25-1230, is a petition for certiorari at the Supreme Court. The Court has called for the views of the Solicitor General; that response is due September 11. Related petitions include Intel v. Squires, No. 26-73, and Tesla v. Granite Vehicle Ventures, No. 26-136. None of those is a holding. A grant, or a change in Office guidance, would reopen parts of what parties are pricing off today.

How we got here
The 2025 sequence is short enough to keep in one place. On February 28, Acting Director Stewart rescinded the Vidal Fintiv memo and treated Sotera as highly relevant but not dispositive. On March 26 the Office bifurcated the process so the Director decides discretion first, and settled expectations became a standing factor — the theme of the SCA Hygiene post from August 2025. On October 17, Director Squires took all institution decisions, began issuing summary notices, and the Office published a same-day NPRM on categorical bars.

The 2026 decisions that matter for consideration are fewer. The March 11 manufacturing and small-business memo added facts the Office will weigh. Tianma, IPR2025-01579, Paper 12 (Mar. 18, 2026) (precedential), held that a foreign government is not a “person” under Return Mail and cannot be petitioner or RPI; “some evidence” can shift the burden, and a conclusory declaration did not carry it in that case. Magnolia, IPR2026-00097, Paper 17 (May 14, 2026) (precedential), is the second-bite case after a district-court miss of the petitioner’s own making. Tesla v. Bulletproof, IPR2026-00204, Paper 14 (June 15, 2026) (informative), collected facts the Office treated as favoring institution — a vacated trial date, a broad stipulation, Office error, young patents, and U.S. manufacturing. It is an informative illustration, not a required template.

Four chairs
On AIA trials, Tesla v. Bulletproof and Magnolia are useful poles rather than a checklist. One shows a record the Office treated as favoring institution; the other shows what happens when the petitioner asks for a second bite after a miss of its own making. After Tianma, foreign-sovereign RPI has to be mapped when it is in play. The working model is current discretion, the proposed bars that are still only proposed, and the stay that may or may not be available on the facts of the parallel case.

Reexam volume is up as petition volume is down. Geotab is the sequencing warning: the IPR was denied on September 12, 2025, and the follow-on reexam (Control No. 90/015,984) was denied on May 22, 2026 under § 325(d). Presentation in an IPR counts even if the Board never reached the merits. Pre-order SNQ paper has been available since April 5, 2026. The proposed RPI identification rule (RIN 0651-AD94) would end anonymity as to the Office; it is still proposed. Chapter 30 is a parallel statute with its own traps. It is not automatically the next filing after a denied IPR.

Prosecution still matters the old-fashioned way. Quality remains the most reliable way off the docket, and after Tesla an examination hole large enough to be called an “Office error” can matter at institution.

Enforcement has to carry more of the weight when an AIA trial is not instituted. District-court filings in the first half of 2026 (1,921, down 18.4%) did not absorb the PTAB contraction, and an IPR filing is not a reliable stay trigger on this docket. That is true under current practice, and it remains true if later guidance changes the institution picture again.

Interested in More?
I will be moderating a panel on this leverage map at the 2026 Midwest IP Institute on Thursday, September 24, 1:15–2:15 p.m., Minnesota CLE Conference Center, 600 Nicollet Mall, Minneapolis: Patent Enforcement in the New Discretionary Denial Era, with Dena Ehrich (3M), Sherry Roberg-Perez (Fortress), Cy Morton (Robins Kaplan), and Jeff Blake (Merchant & Gould). Program. Questions: tbianchi@slwip.com.