Reexam Grant Rates Held Near 95% for a Decade. This Spring They Dropped.

For fourteen years the odds of getting a reexamination ordered barely moved. The USPTO ordered reexamination on about 95% of the requests it decided, year after year, whether the requester was a Fortune 50 defendant or a solo inventor. That stability is the premise behind our explainer on the substantial new question standard, and it held from 2012 straight through the first quarter of 2026, when the grant rate on decided requests was 99%. This spring that changed. In May 2026 the USPTO ordered reexamination on just 70% of the requests it decided, the lowest monthly grant rate in our index going back to 2012.

What the data shows

We track every utility-patent ex parte reexamination and the order that grants or denies each one. Measuring by the date of that order, so that requests still awaiting a decision don't distort the picture, the grant rate had never left the low-to-mid 90s in any full year. Then the spring happened.

May 2026 is the single lowest-grant-rate month in the entire 2012 to 2026 record, and June is the third lowest. Two of the three worst months on record are now back to back. Blended across April through early July, the grant rate is about 85%, roughly ten points under the historical norm. The denial counts tell the same story from the other side: the office has already entered more SNQ denials in the first half of 2026 than in any prior full year in our data, and 2025's full-year total was 21.

Three checks say this is a real shift rather than a data quirk. The denials load into our index within about a day of the order date, so these are freshly issued decisions, not a backlog being cleaned up. They spread across 26 different reexam examiners, so it isn't one examiner running hot. And they hit sophisticated repeat filers, not just thinly argued pro se requests. Among the requesters denied this spring are Apple, Google, Microsoft, Samsung, Tesla, Snap, and Red Hat. When that population starts getting turned away, something in the standard has changed.

Why grant rates are falling

The commentary explains what the counts cannot. Two changes landed on the reexamination front almost simultaneously this spring, and both push in the same direction.

The first is the pre-order procedure that took effect in April 2026. For the first time, a patent owner can file a paper of up to 30 pages, within 30 days of being served with a request, arguing that the request raises no substantial new question, before the office decides whether to order reexamination. Director Squires tied the change directly to volume, citing the recent increase in reexamination requests as the reason for giving patent owners an earlier voice. Morgan Lewis, Cooley, Akin, and Foley all covered the mechanics. Foley's read was that the new paper was unlikely to move the needle "unless it leads to a significant drop in SNQ findings." The spring numbers are that drop. Notably, the pre-order paper is confined to the merits of the substantial new question and cannot raise 325(d) discretion, so this route drives denials on the classic grounds: the art is cumulative, or the question isn't actually new.

The second change runs through 35 U.S.C. 325(d). On May 22, 2026, the USPTO denied a reexamination that Geotab had requested against a Fractus patent, reasoning that the request leaned on substantially the same prior art and arguments as an IPR petition the office had already turned away on discretion. That case sits in our denial data with a May 22 order date, filed less than three months earlier, exactly the window the statute sets for the SNQ decision. Finnegan called it "Discretion All the Way Down." Steptoe called it "the first ex parte reexam domino." Knobbe's Jarom Kesler, speaking to Law360, said it was the "first time I've ever seen this kind of an outcome" across roughly 50 reexams in 15 years, and predicted the office would keep "clamping down on the number of ex parte reexaminations that are going to be allowed." The through-line is that a discretionarily denied IPR no longer clears the way to a reexamination on the same art. It can now foreclose it.

Both mechanisms land in the same bucket we measure. A merits denial under the pre-order procedure and a discretionary denial under 325(d) both show up as a request that never received an order, and both are part of why the spring grant rate fell.

What it means for a request filed today

The practical rule that held for a decade, assume the order will issue and litigate the merits later, no longer runs on autopilot. A few adjustments follow.

Reused IPR art is now a liability, not a shortcut. If the grounds overlap a petition the PTAB denied on discretion, the reexamination request inherits that denial risk. The safer path is genuinely new art or a genuinely new technical light, the kind of question the pre-order paper cannot easily rebut.

The patent owner gets the first word now. A request has to anticipate a 30-page rebuttal aimed squarely at the substantial new question, filed before any order. A request that would have cleared when orders were nearly automatic is not necessarily strong enough to survive that rebuttal.

And the examiner still matters, arguably more than before. The spring denials were spread across more than two dozen examiners, but they were not uniform, and the office has more room than it used to for a discretionary call before ordering. Knowing how the assigned examiner has handled the substantial new question historically is a sharper input now that the base rate is moving.

What to watch

One caveat belongs up front: this is three months of data. The first week of July came back clean, 25 orders and no denials, which is far too small to say whether the spike is easing or the summer will confirm it. The grant rate could revert toward its historical mid-90s, and a single quarter does not make a regime.

But the direction is not coming out of nowhere. Volume is at record highs, the office has said in its own notice that it wants to stem the tide, and it has handed patent owners a new pre-order tool while extending discretionary denial into a forum that never had one. Every firm writing about reexamination this year is describing the same tightening in words. This is what it looks like in numbers. We will update the series as the summer orders land. If the second half of 2026 holds near 85%, the 95% grant rate will have been a fourteen-year fact that ended this spring.

That is the layer we built ExamPat to supply. When the base rate shifts, an aggregate published once a year cannot tell you it is happening, and it cannot tell you how your assigned examiner is deciding the substantial new question right now. Our index can. Check the examiner, and the 325(d) exposure, before you file.