Patents

Who Proves the Diligent Search? Ironburg v. Valve and the Burden of IPR Estoppel

The Federal Circuit adopts a skilled-searcher standard for Section 315(e)(2) estoppel and places the burden of proving it on the patent owner, not the petitioner.

March 16, 2026
Game controller resting on a dark desk beside a keyboard
A dispute over back-paddle buttons on a game controller reset how courts allocate the burden of IPR estoppel. Shutterstock
Educational content, not legal advice. This article explains general legal concepts. It does not create an attorney–client relationship. For your specific situation, consult a licensed attorney.

In Ironburg Inventions Ltd. v. Valve Corp., 64 F.4th 1274 (Fed. Cir. 2023), Nos. 2021-2296, 2021-2297, 2022-1070 (decided April 3, 2023), the U.S. Court of Appeals for the Federal Circuit answered a question that had quietly divided district courts for years: when a defendant who lost an inter partes review later raises invalidity grounds it never put before the Patent Trial and Appeal Board, who bears the burden of showing those grounds were ones it “reasonably could have raised”? Writing for a panel of Judges Lourie, Clevenger (dissenting in part), and Stark, the court held that the estoppel of 35 U.S.C. § 315(e)(2) reaches grounds “a skilled searcher conducting a diligent search reasonably could have been expected to discover” — and, critically, that the burden of proving discoverability falls on the patent owner asserting estoppel, not on the petitioner trying to escape it. The case arose from a willful-infringement judgment out of the Western District of Washington over Ironburg’s U.S. Patent No. 8,641,525, which claims a handheld video game controller with rear-mounted control buttons.

At a glance

  • Case: Ironburg Inventions Ltd. v. Valve Corp., 64 F.4th 1274 (Fed. Cir. 2023), Nos. 2021-2296, 2021-2297, 2022-1070.
  • Decided: April 3, 2023; panel of Lourie, Clevenger (dissenting in part), and Stark.
  • Patent: U.S. Patent No. 8,641,525, directed to a games controller with additional “back” control buttons.
  • Below: A Western District of Washington jury found Valve’s Steam Controller willfully infringed and awarded several million dollars in damages.
  • Two estoppel buckets: “non-petitioned grounds” (never raised in the IPR petition) and “non-instituted grounds” (raised in the petition but not instituted under the pre-SAS partial-institution regime).
  • Holding on non-petitioned grounds: Estoppel applies to grounds a skilled, diligent searcher reasonably could have found; the burden of proof rests on the patent owner. Vacated and remanded because the district court had placed the burden on Valve.
  • Holding on non-instituted grounds: Estoppel applies. A petitioner who did not seek a SAS remand to cure the Board’s partial institution cannot later treat the omitted grounds as untouched by estoppel.

Inter partes review and the estoppel bargain

Inter partes review is the administrative path Congress created in the America Invents Act for a third party to challenge an issued patent’s claims as anticipated or obvious over prior-art patents and printed publications. A petitioner files at the Patent Trial and Appeal Board, the Board decides whether to institute, and an instituted proceeding runs to a written decision that the patent owner can defend and the petitioner can appeal to the Federal Circuit. The attraction is real: a specialized tribunal, a lower burden of proof than the clear-and-convincing standard that governs in district court, and a faster, cheaper record than a jury trial.

That speed comes at a price, and the price is estoppel. Section 315(e)(2) provides that a petitioner who obtains a final written decision in an IPR may not later assert in district court “that the claim is invalid on any ground that the petitioner raised or reasonably could have raised during that inter partes review.” The phrase “raised or reasonably could have raised” is the entire ballgame. “Raised” is easy to police — the grounds are in the petition. “Reasonably could have raised” is the contested frontier, because it reaches grounds the petitioner never actually presented but, in some sense, should have. Ironburg is the Federal Circuit’s most concrete attempt to give that phrase administrable content.

The case reaches the question because Valve, having lost in district court on validity, wanted to keep alive invalidity theories built on prior art it had not used in its earlier IPR. Ironburg invoked estoppel to bar them. The district court agreed with Ironburg in part, but in doing so it made an allocation choice — about who had to prove what — that the Federal Circuit ultimately could not sustain.

The skilled-searcher standard and the burden flip

For the non-petitioned grounds, the panel adopted a discoverability test borrowed from the AIA’s legislative history and from a line of district court decisions: estoppel attaches to any ground “a skilled searcher conducting a diligent search reasonably could have been expected to discover.” The standard is deliberately objective and counterfactual. It does not ask what art the petitioner actually found, or how hard it actually looked. It asks what a hypothetical competent searcher, exercising reasonable diligence, would have turned up. That framing protects patent owners from gamesmanship — a petitioner cannot sandbag obvious prior art and save it for trial — while sparing petitioners from estoppel on genuinely obscure references that no reasonable search would surface.

The more consequential ruling is about burden. Estoppel under § 315(e)(2) is an affirmative defense, raised by the patent owner to knock out the accused infringer’s invalidity case. Following the ordinary rule that the party asserting an affirmative defense must prove it, the court held that the patent owner bears the burden of establishing, by a preponderance of the evidence, that a skilled searcher exercising reasonable diligence would have discovered the ground at issue. The district court had done the opposite, effectively requiring Valve to show that the grounds could not have been found. That was error, and it required vacatur and remand on the non-petitioned grounds so the burden could be placed where it belongs.

The allocation matters more than it might first appear. Discoverability is rarely self-proving; it usually turns on expert testimony about search strategy, classification systems, and how a reference is indexed. Putting that evidentiary load on the patent owner means estoppel is not free — a patentee who wants to foreclose a trial defense must build an actual record about what a diligent search would have produced.

Non-instituted grounds and the long shadow of SAS

The second bucket is a relic of how IPR practice worked before the Supreme Court’s decision in SAS Institute Inc. v. Iancu, 138 S. Ct. 1348 (2018). For a time, the Board would institute review on some petitioned grounds and decline others — partial institution. SAS ended that, holding that the Board must decide the patentability of all claims (and, as the agency read it, all grounds) the petitioner challenges, or none. Valve’s IPR had been instituted in the partial-institution era, so several grounds it actually petitioned were never adjudicated.

Valve argued those non-instituted grounds could not be estopped because they were never part of the instituted proceeding and so were never truly “raised … during” the IPR. The Federal Circuit rejected the argument. After SAS, a petitioner whose proceeding was partially instituted could move to have the Board take up the omitted grounds; many did. Valve did not. The court held that a petitioner’s “choice to leave unremedied the Board’s mistake” does not shield the omitted grounds from estoppel — having put the grounds in its petition and then declined the available cure, Valve could not treat them as if they had never been raised. Estoppel applied. (Judge Clevenger parted ways with the majority on a separate indefiniteness question concerning the claim term describing a button extending “substantially the full distance” of the controller, not on the estoppel analysis.)

Open questions

  • How much search evidence is enough? The opinion sets the standard but not the threshold. Lower courts must still decide what quantum of expert testimony about classification, databases, and search terms satisfies a patent owner’s burden.
  • Does the skilled searcher know the patent’s prosecution? The hypothetical searcher’s starting knowledge — the field, the claims, the cited art — is unspecified, and that baseline can swing whether a reference was “reasonably” discoverable.
  • What about system or product prior art? Section 315(e)(2) estoppel is often argued to be limited to patents and printed publications, the only art available in an IPR; Ironburg does not resolve how the skilled-searcher test interacts with device-based invalidity grounds that an IPR could never have reached.
  • Will non-instituted grounds keep appearing? The pre-SAS posture is fading, but cases instituted in that window continue to surface, and the court’s “you should have sought a remand” reasoning will govern them.

Implications

  • Patent owners must build a discoverability record. Invoking estoppel is no longer a citation to § 315(e)(2); it is an evidentiary undertaking requiring proof of what a diligent search would have found.
  • Petitioners gain a real defense. A petitioner can defeat estoppel by showing a reference was genuinely obscure — but the patent owner now has to open that issue first.
  • Search documentation has new value. Contemporaneous records of prior-art searches, on both sides, become litigation evidence about reasonable diligence.
  • Petition drafting is a strategic estoppel decision. Every ground included is plainly “raised”; every reasonably discoverable ground omitted is at risk of being estopped anyway, narrowing the upside of a slimmed-down petition.
  • Old IPRs still bite. Defendants whose IPRs predate SAS should assume non-instituted grounds are estopped unless they pursued a remedial path at the time.

Frequently asked questions

What is the “skilled searcher” standard? It is an objective test for IPR estoppel: § 315(e)(2) bars invalidity grounds that a hypothetical skilled searcher, conducting a diligent search with reasonable effort, could have been expected to discover — regardless of whether the petitioner actually found or used them.

Who has to prove that a ground was discoverable? After Ironburg, the patent owner does. Estoppel is an affirmative defense, so the patentee must show by a preponderance of the evidence that a diligent search would have surfaced the ground the accused infringer now wants to assert.

Why are non-instituted grounds estopped if the Board never ruled on them? Because the petitioner put them in its petition and, after SAS Institute v. Iancu abolished partial institution, could have moved to have the Board adjudicate them. A petitioner that declined that cure cannot later claim the grounds were never raised.

Authorities and sources