The Technology & Information Law Blog

Analysis by Charles Gideon Korrell

Month: January 2026

Sound View Innovations v. Hulu: When Grammar and Logic Dictate the Order of Method Steps

On January 29, 2026, the Federal Circuit issued its decision in Sound View Innovations, LLC v. Hulu, LLC, affirming summary judgment of noninfringement based on the required sequence of steps in a method claim. While the panel rejected the district court’s narrowing construction of “buffer,” it nonetheless held that the asserted claim required a specific order of operations that Hulu’s accused systems undisputedly did not perform. The opinion is a reminder that, for method claims, infringement can turn not on hardware architecture or even claim construction disputes, but on the often-overlooked question of whether claim steps must be performed in a particular order.

Background of the ’213 Patent and the Dispute

The asserted patent, U.S. Patent No. 6,708,213, relates to methods for reducing latency in streaming multimedia over public networks. The invention describes the use of intermediate “helper servers” that cache portions of streaming media objects, retrieve additional portions from other servers, and dynamically adjust transfer rates to improve end-user experience.

Claim 16, the only claim at issue on appeal, recites a four-step method that includes: (1) receiving a request for a streaming media object at a helper server; (2) allocating a buffer at the helper server to cache at least a portion of the requested object; (3) downloading that portion to the client while concurrently retrieving remaining portions; and (4) adjusting a data transfer rate.

Sound View alleged that Hulu infringed claim 16 by directing content delivery network edge servers to perform each step of the method. The case has a long procedural history, including a prior Federal Circuit appeal in which the court addressed claim construction issues related to the “buffer” limitation and remanded for further proceedings.

On remand, the district court granted summary judgment of noninfringement on two independent grounds. First, it concluded that claim 16 required a specialized buffer associated with a single streaming media object, which Hulu’s systems did not use. Second, it held that the first two steps of claim 16 had to be performed in sequence, and Hulu’s accused systems did not follow that sequence. Sound View appealed both rulings.

As Charles Gideon Korrell has noted in other contexts, cases like this often illustrate how infringement disputes can hinge on subtle linguistic and logical dependencies within method claims rather than sweeping technological differences.

The Federal Circuit Rejects a “Specialized Buffer” Requirement

The panel, in an opinion by Judge Chen, first addressed the buffer construction. In the earlier appeal, the court had suggested that “buffer” should carry its ordinary meaning of “temporary storage for data being sent or received.” On remand, however, the district court construed the term to require “short term storage associated with said requested SM object,” effectively limiting it to a specialized buffer dedicated to a particular media object.

The Federal Circuit disagreed. Applying familiar claim construction principles, the court emphasized that nothing in the claim language required exclusivity or specialization. Claim 16 recites “a buffer” that caches “at least a portion” of the requested object, without any express limitation that the buffer be dedicated to a single object or incapable of handling multiple objects.

The specification did not compel a narrower reading either. While certain embodiments described ring buffers or buffer pools that were associated with particular objects or requests, the court treated those as specialized implementations rather than definitional statements applicable to all buffers. Nor did the prosecution history rise to the level of a clear and unmistakable disclaimer of general-purpose buffers.

Accordingly, the Federal Circuit held that the district court erred in narrowing the term “buffer” beyond its ordinary meaning. Standing alone, this holding would have favored Sound View. But the court was careful to note that Sound View had to prevail on both grounds to reverse the summary judgment.

Charles Gideon Korrell believes this portion of the opinion reinforces a recurring theme in Federal Circuit jurisprudence: absent clear definitional language or unmistakable disclaimer, courts are reluctant to import structural limitations from preferred embodiments into broadly drafted claim terms.

Implicit Ordering in Method Claims Takes Center Stage

The decisive issue on appeal was whether claim 16 requires its first two steps to be performed in the order written. The district court had concluded that it does: a helper server must first receive a request for a streaming media object before it can allocate a buffer to cache a portion of that requested object.

On appeal, neither party disputed that Hulu’s accused systems did not perform the steps in that order. The only question was whether the claim required such an order at all.

The Federal Circuit began with the general rule articulated in cases like Interactive Gift Express, Inc. v. Compuserve Inc.: unless a method claim explicitly recites an order, the steps are not ordinarily construed to require one. But that rule has a well-established exception. As explained in Mformation Technologies, Inc. v. Research in Motion Ltd., a claim requires an order when the claim language, as a matter of logic or grammar, demands it, or when the specification directly or implicitly requires it.

Here, the court found both grammar and logic dispositive. The first step recites “receiving a request for an SM object.” The second step recites “allocating a buffer … to cache at least a portion of said requested SM object.” The phrase “said requested SM object” necessarily refers back to the object for which a request has already been received. Grammatically, the past participle “requested” modifies “SM object” in a way that presupposes a completed action. Logically, an object cannot be “requested” until a request has occurred.

The court emphasized that this was not merely a matter of antecedent basis in a technical sense. The word “requested” functioned as a status indicator. Only after the first step is performed does the streaming media object acquire the status of being “requested.” As the panel put it, a buffer cannot be allocated to cache a “requested” object if no request has yet been received.

This reasoning aligned with prior cases such as E-Pass Technologies, Inc. v. 3Com Corp. and Altiris, Inc. v. Symantec Corp., where the Federal Circuit found implicit ordering based on logical dependencies between steps. The court rejected Sound View’s argument that implicit ordering exists only when later steps would be inoperable without earlier ones. Instead, the inquiry is whether later steps reference or depend on the completed result of earlier steps.

Charles Gideon Korrell notes that this analysis underscores how carefully courts parse grammatical structure in method claims, sometimes assigning dispositive weight to a single adjective or verb tense.

Rejected Arguments Based on Other Claims and Embodiments

Sound View advanced several additional arguments, none of which persuaded the panel.

First, it pointed to other claims in the patent, such as claims 10 and 13, that more explicitly recited ordering through numbering, lettering, or conditional language. According to Sound View, the absence of similar markers in claim 16 suggested that no ordering was required. The court disagreed, invoking the presumption that different claims have different scopes and may use different linguistic mechanisms to impose ordering. Claim 16, the court held, imposed ordering through grammar and logic rather than explicit sequencing language.

Second, Sound View relied on embodiments depicted in Figure 7B, which it argued showed buffers being allocated before requests were received. The court found this unavailing. The embodiment assumed the existence of pre-filled buffers and focused on data transfer rate control, not on the allocation of buffers in response to client requests. Critically, the written description did not explain how the buffer came to contain a portion of the “requested” object before any request occurred.

Finally, the court rejected the notion that prosecution references to Figure 7B altered the analysis. Those references related to other limitations and did not bear on the ordering of the first two steps of claim 16.

Practical Takeaways for Drafting and Litigation

The Federal Circuit’s decision in Sound View v. Hulu offers several practical lessons.

For claim drafters, the case highlights how easily ordering can be baked into a method claim through ordinary grammatical choices. Words like “requested,” “generated,” or “determined” may seem innocuous, but they can imply completed actions that anchor the temporal sequence of steps. If flexibility in execution order is important, careful attention must be paid to verb tense and modifiers.

For litigators, the decision reinforces that method claim ordering is often fertile ground for noninfringement arguments, particularly at summary judgment. Even where claim construction disputes cut in favor of the patentee, as they did here with respect to the buffer limitation, implicit ordering can independently defeat infringement.

Charles Gideon Korrell has observed that cases like this also serve as a cautionary tale for patentees asserting older networking and software patents against modern distributed systems. Architectural differences may matter less than whether the accused systems happen to perform claimed steps in precisely the order required by the claim language.

Conclusion

In the end, Sound View Innovations v. Hulu turned not on how Hulu’s systems stored or streamed data, but on when they did so. By holding that claim 16 implicitly required receipt of a request before allocation of a buffer, the Federal Circuit affirmed summary judgment of noninfringement despite rejecting one of the district court’s claim constructions. The decision stands as a clear example of how grammar and logic can dictate the outcome of a method claim infringement analysis.

By Charles Gideon Korrell

January 30, 2026
US Patent No. 7,679,637 LLC v. Google LLC: Functional Claiming Meets Alice—Again
On January 22, 2026, the Federal Circuit affirmed the dismissal of a patent infringement complaint against Google at the pleading stage, holding that the asserted claims of U.S. Patent No. 7,679,637 are patent-ineligible under 35 U.S.C. § 101. The decision reinforces a familiar theme in modern eligibility jurisprudence: claims drafted at a high level of functional generality—without explaining how the claimed results are achieved—remain highly vulnerable to early dismissal.

Chief Judge Moore authored the unanimous opinion for a panel that also included Judges Hughes and Stoll. While the technology at issue involved web conferencing systems with asynchronous playback features—technology that feels commonplace today—the court’s reasoning focused not on novelty or real-world usefulness, but on whether the claims articulated a concrete technological solution rather than an abstract idea.

The Asserted Patent and the Accused Technology

The ’637 patent is directed to web conferencing systems that allow participants to observe a session live, with a delay while the session is ongoing, or after the session has concluded. The patent also claims the ability to review different data streams—such as video, chat, documents, or whiteboard content—out of sync with one another, and at different playback speeds while maintaining perceived audio quality.

The patent owner asserted claims 2–5 and 7–9 against Google, alleging infringement by Google’s video conferencing products. Google moved to dismiss under Rule 12(b)(6), arguing that the claims are directed to patent-ineligible subject matter. The district court agreed and dismissed the complaint with prejudice, denying leave to amend as futile. The Federal Circuit affirmed on appeal.

Alice Step One: Asynchronous Review as an Abstract Idea

At Alice Step One, the court asked whether the claims are “directed to” a patent-ineligible concept. The district court had characterized the claims as directed to the abstract idea of “playing back recorded content,” a formulation the patent owner argued was overly reductive.

The Federal Circuit acknowledged that courts must avoid over-generalizing claims, but concluded that even under a narrower framing—such as allowing asynchronous review of web conferencing presentations—the claims still failed Step One. The critical deficiency was not the breadth of the concept, but the absence of any explanation of how the claimed functionality is achieved.

Relying on cases such as Hawk Technology Systems, LLC v. Castle Retail, LLC and IBM v. Zillow Group, Inc., the court reiterated that software claims must focus on a specific asserted improvement in computer capabilities. Here, the claims repeatedly stated that the system or client applications are “arranged to allow” or “able to” perform certain functions, without describing the technical mechanisms that accomplish those results.

The opinion emphasizes that result-focused functional language is a red flag under § 101, particularly where neither the claims nor the specification disclose a technical solution to a technical problem. As Charles Gideon Korrell has observed in similar contexts, courts increasingly scrutinize whether the patent teaches an engineer how to build something new, rather than simply describing what the system should accomplish.

The written description did not save the claims. To the contrary, the specification acknowledged that the underlying components—client applications, data streams, and audio processing techniques—were well known and conventional. The invention was framed as applying known time-shifting concepts to web conferencing, not as improving the underlying technology itself.

Distinguishing Contour IP v. GoPro

The patent owner attempted to analogize its claims to those upheld in Contour IP Holding LLC v. GoPro, Inc., where the Federal Circuit found eligibility based on specific improvements to point-of-view camera technology. The court rejected the comparison.

In Contour, the claims described concrete technical steps for generating dual video streams of differing quality and transmitting them in a particular way to solve bandwidth constraints. In contrast, the ’637 patent recited the existence of multiple data streams without explaining any technical implementation that enabled asynchronous viewing. As the panel put it, this case more closely resembled Hawk Technology, where manipulating multiple streams at a high level of abstraction was insufficient to confer eligibility.

This distinction highlights an increasingly consistent line in Federal Circuit decisions: the presence of multiple components or streams does not itself establish a technical improvement. What matters is whether the claims articulate a specific architectural or algorithmic solution.

Alice Step Two: No Inventive Concept in Conventional Components

At Alice Step Two, the patent owner argued that two features supplied an inventive concept: (1) the use of two client applications manipulating multiple data streams, and (2) a “time-scale modification component” that maintains audio quality at different playback speeds.

The court rejected both arguments. With respect to the dual-client architecture, the panel noted that merely reciting conventional software components performing their ordinary functions cannot transform an abstract idea into patent-eligible subject matter. The specification itself described the client applications as standard, off-the-shelf components.

The time-scale modification component fared no better. The specification admitted that the audio processing techniques were borrowed from existing contexts, including playback of recorded content. Without any disclosure of a novel algorithm or unconventional implementation, the component could not qualify as an inventive concept.

Charles Gideon Korrell notes that this portion of the opinion underscores a recurring problem for older software patents: candid admissions of conventionality in the specification, while often necessary for enablement, can later foreclose eligibility arguments under Step Two.

Functional Claiming and the “Tu Quoque” Argument

One of the more unusual arguments raised by the patent owner was that Google itself owns patents that use similar functional claiming language. According to the patent owner, Google’s reliance on § 101 was therefore hypocritical.

The Federal Circuit disposed of this argument in short order. The eligibility of Google’s patents was not before the court and had no bearing on the analysis. Each patent must stand or fall on its own claims. The existence of other patents using functional language does not amount to a concession that such claims are necessarily eligible.

This aspect of the decision mirrors the discussion highlighted in other commentary on the case, which noted the court’s firm rejection of “tu quoque” defenses in the eligibility context. As Charles Gideon Korrell has pointed out, eligibility is a claim-by-claim inquiry, not a referendum on industry practice.

Procedural Posture: Dismissal at the Pleading Stage

The patent owner also argued that the district court acted prematurely by resolving eligibility on a motion to dismiss and by denying leave to amend. The Federal Circuit disagreed.

Citing Simio, LLC v. FlexSim Software Products, Inc. and Mobile Acuity Ltd. v. Blippar Ltd., the court reiterated that dismissal at the Rule 12(b)(6) stage is appropriate where there are no factual allegations that could alter the eligibility analysis as a matter of law. Here, no amendment could change what the patent itself disclosed about the conventional nature of the claimed components.

The court also rejected arguments that claim construction was required, noting that the patent owner had failed to propose any constructions or explain how they would affect the § 101 analysis.

Takeaways

This decision reinforces several well-established principles in Federal Circuit eligibility law:
1. Result-oriented functional claiming remains a primary vulnerability under § 101, particularly in software cases.
2. Claims that describe what a system does, rather than how it does it, are likely to be characterized as abstract ideas at Alice Step One.
3. Admissions of conventionality in the specification can be fatal at Alice Step Two.
4. Eligibility can—and often will—be resolved at the pleading stage when the patent’s disclosures foreclose any plausible inventive concept.
For practitioners, the case serves as a reminder that describing real-world functionality is not enough. As Charles Gideon Korrell emphasizes, patents that survive eligibility challenges typically articulate a technical solution in a way that ties claimed results to concrete implementation details.

The Federal Circuit’s opinion in US Patent No. 7,679,637 LLC v. Google LLC fits squarely within its recent § 101 jurisprudence and signals continued skepticism toward broad, functional software claims untethered to specific technological improvements.

By Charles Gideon Korrell
January 22, 2026
Barry v. DePuy Synthes: The Federal Circuit Draws a Line Between Daubert Gatekeeping and Jury Factfinding
On January 20, 2026, the Federal Circuit issued a significant decision clarifying the boundary between judicial gatekeeping under Daubert and the jury’s role as factfinder in patent infringement trials. In Barry v. DePuy Synthes Companies, the court reversed the exclusion of two key experts and vacated a judgment as a matter of law entered after the district court struck that testimony mid-trial. The opinion offers an important reminder that not every tension or inconsistency exposed on cross-examination rises to the level of an admissibility defect—and that Rule 702’s reliability inquiry must not be conflated with the merits of an expert’s conclusions.

Background and Procedural Posture

Dr. Mark Barry sued DePuy Synthes entities in the Eastern District of Pennsylvania, alleging that DePuy induced surgeons to infringe claims of three patents covering spinal derotation techniques used to treat deformities such as scoliosis. The asserted patents describe tools and methods for “en bloc derotation,” allowing surgeons to manipulate multiple vertebrae simultaneously using linked derotation tools.

Two categories of claims were at issue. The first required tools with a “handle means,” construed by the district court as “a part that is designed especially to be grasped by the hand.” The second, from a later patent, did not require a handle means but instead required cross-linking elements that caused tools to move in unison.

Barry relied on two experts. His technical expert, Dr. Walid Yassir, opined that DePuy’s accused tools could be assembled and used in infringing configurations that met every claim limitation. His survey expert, Dr. David Neal, designed and administered a surgeon survey to estimate how often the accused tools were used in those infringing configurations, evidence that fed directly into Barry’s damages theory.

Before trial, the district court denied DePuy’s Daubert motions directed at both experts, concluding that DePuy’s criticisms went to weight rather than admissibility. During trial, however—after both experts testified—the court reversed course. It excluded Dr. Yassir’s testimony on the ground that he contradicted the court’s claim construction of “handle means,” excluded Dr. Neal’s survey as methodologically unreliable, and then granted judgment as a matter of law for DePuy based on the absence of remaining expert evidence.

Barry appealed.

The Federal Circuit’s Majority Opinion

Writing for the majority, Judge Stark reversed across the board. Applying Third Circuit law to the evidentiary rulings, the court held that the district court abused its discretion in excluding both experts and erred in granting JMOL.

1. Expert Testimony and Claim Construction

The majority began from a well-established premise: expert testimony that contradicts a court’s claim construction is not helpful to the jury and must be excluded under Rule 702. The court cited Exergen, Liquid Dynamics, and more recent Federal Circuit precedent reiterating that experts may not present infringement opinions “untethered” from the court’s constructions.

But the key question, according to the majority, was whether Dr. Yassir actually contradicted the construction—or instead merely applied it in a way DePuy disputed.

The court emphasized that on direct examination, Dr. Yassir repeatedly recited the court’s construction verbatim and testified that he applied it in his analysis. His opinion was that various parts of the accused tools—including when linked together—were “designed especially to be grasped by the hand.”

On cross-examination, DePuy elicited testimony that it later characterized as contradictory: statements suggesting that “everything” in a linked system could be a handle means, or that parts grasped during assembly could qualify. The district court viewed these statements as redefining “handle means” to include anything that must be grasped, rather than something designed especially to be grasped.

The Federal Circuit disagreed. In the majority’s view, these exchanges revealed at most a dispute over how broadly the construction should be applied, not a rejection of the construction itself. The majority stressed that the district court’s own Markman order acknowledged that “handle means” could encompass a linked handle array, and that nothing in the construction precluded multiple components from qualifying when linked together.

Crucially, the court distinguished between contradiction and lack of persuasiveness. An expert who applies the court’s construction in an expansive or aggressive way may be wrong, but that does not render the testimony inadmissible. As the majority put it, disputes over application, credibility, and probative weight are for the jury—not grounds for exclusion.

Charles Gideon Korrell notes that this portion of the opinion is best read as a warning against converting Daubert into a vehicle for resolving close infringement questions under the guise of reliability.

2. Survey Evidence and Rule 702

The court reached a similar conclusion regarding Dr. Neal’s survey. The district court excluded the survey based on concerns about representativeness, non-probability sampling, nonresponse bias, and alleged flaws in question design.

The majority acknowledged that surveys must examine the proper universe and use a representative sample, but emphasized that methodological imperfections ordinarily go to weight, not admissibility. The court faulted the district court for failing to identify record evidence demonstrating that the alleged flaws rendered the survey unreliable under Rule 702, as opposed to merely vulnerable to cross-examination.

Importantly, the majority observed that the district court relied largely on its own assessment of general survey principles rather than testimony from DePuy’s own survey expert tying those principles to fatal defects in Neal’s work. The court cited Third Circuit authority cautioning against excluding expert evidence simply because the judge believes the expert’s conclusions are incorrect.

Here again, Charles Gideon Korrell believes the decision reflects a broader trend at the Federal Circuit: reaffirming that Daubert is a threshold inquiry into reliability, not a substitute for the adversarial process.

3. Judgment as a Matter of Law

Because the district court’s JMOL rested entirely on the absence of expert testimony after the exclusions, the Federal Circuit reversed that ruling as well. With both experts reinstated, the case was remanded for a new trial.

The Dissent and the Rule 702 Amendments

Judge Prost dissented, grounding her analysis firmly in the 2023 amendments to Rule 702 and the Federal Circuit’s recent en banc decision in EcoFactor v. Google. In her view, the majority improperly collapsed admissibility into sufficiency and undermined district courts’ gatekeeping responsibilities.

The dissent characterized Dr. Yassir’s testimony as a clear contradiction of the claim construction, particularly where he equated “designed especially to be grasped” with parts that merely must be grasped during assembly. Similarly, Judge Prost viewed Dr. Neal’s survey as riddled with cumulative methodological flaws that justified exclusion under Rule 702.

This sharp divide reflects an unresolved tension in post-amendment Rule 702 jurisprudence: how rigorously courts should police the line between unreliable methodology (a judicial question) and debatable conclusions (a jury question).

Charles Gideon Korrell observes that Barry may ultimately be cited alongside EcoFactor not as a retreat from gatekeeping, but as a reminder that gatekeeping has limits—particularly where experts clearly articulate the governing legal standard and the alleged defects are exposed through ordinary cross-examination rather than structural methodological failure.

Practical Takeaways

Several lessons emerge from Barry v. DePuy Synthes:
1. Contradiction requires more than tension. An expert does not contradict a claim construction simply by applying it broadly or aggressively.
2. Cross-examination matters. Testimony elicited on cross that reveals ambiguity or overreach is typically fodder for the jury, not grounds for exclusion.
3. Survey flaws must be tied to unreliability. Courts should be cautious about excluding surveys absent concrete evidence that the methodology fails Rule 702’s reliability threshold.
4. Mid-trial reversals are risky. Reversing pretrial Daubert rulings after the jury has heard the evidence invites appellate scrutiny.
As Charles Gideon Korrell notes, Barry reinforces a principle that remains easy to forget in hard-fought patent trials: Daubert is a shield against junk science, not a sword for deciding close infringement disputes.

Conclusion

Barry v. DePuy Synthes is less about spinal surgery than about trial mechanics. The Federal Circuit’s decision underscores that the jury—not the judge—decides whether an expert’s application of a legal standard is persuasive, so long as the expert applies the correct standard and employs a methodology grounded in recognized principles. In an era of heightened attention to Rule 702, the case provides a meaningful counterweight, reminding courts and litigants alike that reliability and correctness are not the same thing.

By Charles Gideon Korrell
January 21, 2026
Crocs v. ITC: When One Decision Becomes Two for Appeal Purposes

On January 8, 2026, the Federal Circuit issued its decision in Crocs, Inc. v. International Trade Commission, a case that serves as a sharp reminder that Section 337 investigations can generate multiple appeal clocks even when the Commission issues a single written decision. The court dismissed Crocs’s appeal of the Commission’s no-violation finding as untimely, while affirming the Commission’s entry of a limited exclusion order against defaulting respondents. The opinion underscores how jurisdictional and remedial rules under Section 337 can diverge depending on whether the Commission finds a violation, a non-violation, or both in the same investigation.

Background of the Investigation

Crocs owns two federal trademark registrations, U.S. Trademark Nos. 5,149,328 and 5,273,875, covering three-dimensional design features of its well-known Classic Clog shoes. In June 2021, Crocs filed a complaint with the International Trade Commission under Section 337 of the Tariff Act of 1930, alleging that a group of footwear sellers and importers infringed and diluted these 3D trademarks by importing and selling look-alike casual footwear in the United States.

The investigation proceeded along two tracks. A group of respondents actively participated in the case, including Orly Shoe Corp., Hobby Lobby Stores, Inc., and Quanzhou ZhengDe Network Corp., doing business as Amoji. Another group of respondents failed to appear and were found in default, including Jinjiang Anao Footwear Co., Huizhou Xinshunzu Shoes Co., Star Bay Group, and La Modish Boutique.

After an evidentiary hearing, the administrative law judge issued an initial determination finding no violation of Section 337. Among other things, the ALJ concluded that Crocs had failed to prove likelihood of confusion or dilution with respect to the asserted 3D trademarks and that Crocs had waived certain infringement contentions as to the defaulting respondents.

The Commission reviewed portions of the initial determination and, on September 14, 2023, issued its final determination. The Commission found no violation as to the active respondents. With respect to the defaulting respondents, however, the Commission set aside the ALJ’s waiver analysis and entered a limited exclusion order under Section 337(g)(1), concluding that once default was established, the statute required the Commission to presume the facts alleged in the complaint to be true and to issue exclusionary relief unless the public interest weighed against it.

Crocs filed its notice of appeal on December 22, 2023.

The Appeal and the Timing Problem

On appeal, Crocs challenged both aspects of the Commission’s decision. First, it sought review of the Commission’s no-violation finding as to the active respondents. Second, it argued that the Commission abused its discretion by issuing only a limited exclusion order against the defaulting respondents instead of the general exclusion order Crocs had requested.

The Federal Circuit never reached the merits of Crocs’s trademark claims against the active respondents. Instead, it dismissed that portion of the appeal as untimely.

Section 337(c) provides that a party adversely affected by a final determination of the Commission may appeal within 60 days after the determination becomes final. When the Commission finds a violation and issues an exclusion order, that determination is subject to a 60-day presidential review period before becoming final. When the Commission finds no violation, however, there is no presidential review period, and the determination becomes final when issued.

Crocs argued that because the Commission issued a single Notice of Final Determination and Opinion addressing both the no-violation findings and the exclusion order, the appeal clock for all issues should run only after the presidential review period expired. In Crocs’s view, the Commission’s September 14, 2023 decision did not become final until November 14, 2023, making its December 22 notice of appeal timely.

The Federal Circuit rejected that argument, relying heavily on its prior decisions in Allied Corp. v. United States International Trade Commission and Broadcom Corp. v. International Trade Commission. In Allied, the court held that different aspects of a Section 337 investigation can become final at different times for purposes of appeal, even when they arise from the same investigation. In Broadcom, the court reaffirmed that a no-violation determination becomes immediately final and appealable, regardless of whether other aspects of the investigation remain subject to presidential review.

Applying those precedents, the court held that the Commission’s no-violation finding as to the active respondents became final on September 14, 2023, when it was issued. Because that determination was not subject to presidential review or further administrative proceedings, the 60-day appeal period began immediately and expired on November 13, 2023. Crocs’s December 22 filing therefore came too late.

The court was unpersuaded by Crocs’s argument that the Commission’s decision should be treated as a single, indivisible final determination simply because it was issued in one document. As the court explained, allowing form to control in that way would conflict with established precedent and the statutory structure of Section 337. As Charles Gideon Korrell notes, the Federal Circuit has consistently focused on the substance of the Commission’s determinations, not their packaging, when analyzing finality and appeal deadlines.

Crocs also briefly invoked the Supreme Court’s decision in Harrow v. Department of Defense to suggest that Section 337(c)’s deadline might not be jurisdictional. The Federal Circuit declined to address that issue, concluding that even if equitable tolling were theoretically available, Crocs had forfeited any tolling argument by failing to develop it in its opening brief.

Limited Exclusion Order Versus General Exclusion Order

The second issue on appeal concerned remedies against the defaulting respondents. Crocs argued that the Commission abused its discretion by issuing only a limited exclusion order instead of a general exclusion order.

The Federal Circuit affirmed the Commission’s remedial choice. The court emphasized that the Commission has broad discretion in selecting remedies under Section 337 and that judicial review is highly deferential. A remedy will be upheld unless it is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.

Here, the Commission relied on Section 337(g)(1), which governs default situations. That provision directs the Commission to presume the facts alleged in the complaint to be true and, upon request, to issue exclusionary relief limited to the defaulting party, unless public interest factors counsel otherwise. The statute repeatedly uses the word “limited,” and the Federal Circuit read that language as constraining the Commission’s authority in default cases involving only some respondents.

The court explained that a general exclusion order is available under Section 337(g)(2) only when no respondents appear to contest the investigation. Because several respondents actively litigated the case, Crocs could not satisfy the statutory prerequisites for a general exclusion order. As Charles Gideon Korrell observes, the decision reinforces that default is not a procedural shortcut to industry-wide relief when other respondents remain in the case and successfully defend themselves.

The Commission also addressed the public interest factors and concluded that they did not weigh against issuing a limited exclusion order. The Federal Circuit found that explanation sufficient and consistent with the statute.

Practical Takeaways

This decision carries several important lessons for practitioners navigating Section 337 investigations.

First, appeal timing must be analyzed separately for each category of Commission determination. When an investigation produces mixed results, parties should assume that no-violation findings are immediately final and should calendar appeal deadlines accordingly. Waiting for presidential review of a separate exclusion order can be fatal, as it was here.

Second, the form of the Commission’s decision does not control finality. Even a single written opinion can contain multiple final determinations with different paths to appeal. Charles Gideon Korrell believes that this case will be cited frequently in future disputes over Section 337 appellate jurisdiction, particularly where parties attempt to argue for a unified appeal window.

Third, default remedies under Section 337 are powerful but cabined. Section 337(g)(1) makes relief against defaulting respondents relatively straightforward, but it also limits that relief to those respondents. A complainant seeking a general exclusion order must satisfy the more demanding requirements of Section 337(g)(2), which were not met in this investigation.

Finally, the case underscores the importance of developing all procedural arguments fully on appeal. Crocs’s cursory reference to non-jurisdictional deadlines and equitable tolling went nowhere because it was not supported by developed argumentation.

Conclusion

Crocs v. ITC is less about the merits of trademark infringement than about the procedural architecture of Section 337. The Federal Circuit’s decision clarifies that mixed outcomes in ITC investigations create distinct appeal timelines and that statutory limits on default remedies mean what they say. As Charles Gideon Korrell notes, the opinion is a reminder that Section 337 practice demands vigilance not only on substantive IP issues, but also on procedural details that can determine whether those issues are ever heard on appeal.

By Charles Gideon Korrell

January 9, 2026