Federal Circuit Wrestles with Prosecution Laches in Sonos v. Google

Patent

In 2020, Sonos sued Google in the N.D. of California, asserting two “zone scenes”  wireless speaker patents (US 10,848,885 and 10,469,966) that allow for overlapping groups of speakers. A jury found Google liable and awarded $30 million in damages. In an unusual move, District Judge William Alsup threw out the verdict in a 55-page post-trial order. He held the patents unenforceable for prosecution laches and invalid for lack of written description support. Judge Alsup found Sonos had engaged in a “daisy chain” continuation strategy to keep the patent family alive for over 13 years.  Although longer than average, this timeline is not long for a patent family recognized as valuable.  But, one quirk is that the particular “overlapping zone scene” claims  were not presented in claim form until 2019. And, they were added after Google began selling products with that feature. Judge Alsup also concluded that Sonos’s 2019 amendments were directed to new matter not supported by the original 2006–2007 applications — i.e., lacking written description suport.  Without that priority claim, those earlier filings became anticipatory prior art to the later would-be continuation.

The appeal is now pending, and the Federal Circuit held oral arguments on July 10, 2025.  The case raises fundamental questions about the scope of the “sparingly applied” doctrine of prosecution laches and also about the proper procedure for raising written description/anticipation issues during trial.

I believe that a typical Federal Circuit panel would side with the appellant Sonos. But, the panel in this case is somewhat unique. The combination of Judges Prost and Lourie creates an interesting dynamic given their opposing positions in the 2010 prosecution laches case of Cancer Research. Judge Prost’s dissent in that case shows her willingness to apply prosecution laches.  Although Judge Lourie did not find laches in Cancer Research, he is Judge Lourie authorship of Symbol v. Lemelson—the decision that revived prosecution laches in 2005.

In ratings of “patent friendliness” (including one by Google’s attorney Dan Bagatel), Judges Prost and Lourie are among the three least patent-friendly judges on the Federal Circuit.  Judge Bumb, sitting by designation from the District of New Jersey, adds unpredictability as she lacks the extensive patent law track record of her colleagues. All three are Republican appointees who generally embrace textualist approaches, but that judicial philosophy matters less here since prosecution laches is an entirely judge-made equitable doctrine with minimal congressional guidance beyond Section 282’s broad authorization of equitable defenses.

Sonos’s counsel, Josh Rosenkranz is a partner at Orrick and heads its Supreme Court & Appellate Practice. A former clerk to both Justice William J. Brennan Jr. and Judge Antonin Scalia, he has argued a large number of cases, including many patent cases.  Dan Bagatell represents Google and is a partner at Perkins Coie where he chairs the firm’s Federal Circuit Patent Appeals Practice.

Rosenkranz framed Judge Alsup’s laches holding as an unprecedented threat to routine continuation practice, while Bagatell defended it as a rare but justified response to egregious delay. The oral arguments focused on this tension as well as the unusual procedural history that included Judge Alsup’s sua sponte revival of the written description issue during trial, and his “shootout” summary judgment approach that led to some ongoing confusion.

Endangering Continuation Practice:  Judge Alsup’s decision was remarkable because it seems to implicate broad swath’s of standard continuation practice. Rosenkranz, opened by asserting that the district court had overextended prosecution laches to a degree that threatens standard patent continuation practice. “The district court took a doctrine that this court has cautioned must be sparingly applied and stretched it to endanger standard continuation practice.” If allowed to stand, he warned, Judge Alsup’s approach “will fundamentally alter the norms of both patent prosecution and patent litigation.”

Rosenkranz then enumerated three “undisputed” facts that, in Sonos’s view, undercut any laches finding:

  1. Diligent Prosecution: Judge Alsup himself found that Sonos had “diligently prosecuted patent applications in this family” throughout the 13-year period. Sonos never sat on its rights; it kept the “daisy chain” alive with continuous filings.
  2. PTO-Caused Delay: More than half of the total delay was due to Patent Office backlog, not Sonos. “Seven years [of the delay] was attributable to the PTO,” leaving at most “a six-year delay” by the applicant.
  3. Early Broad Claims: Sonos had already obtained earlier patents with broader claims covering all “zone scenes” (with no limitation on overlapping zones) “well before Google launched its products.” In other words, by 2013–2014 Sonos had patent coverage of the general zone scene technology (indeed Sonos points to its 2014 ’228 patent) which put Google on notice. Sonos was not lurking in the weeds with a secret patent; it had disclosed its technology and warned Google in 2016 of potential infringement.

Rosenkranz emphasized what was “missing from this case” as compared to classic prosecution laches scenarios: “any of the shenanigans” typically present in laches cases. Sonos never abandoned any application in the chain to manipulate filing dates, nor did it “bury the patent office” in duplicative filings or engage in other bad-faith delay tactics. In short, Sonos’s use of successive continuations was within normal practice – not an abuse warranting the draconian remedy of unenforceability.  This case is also different from prior laches cases because of the 20 year patent term — delays in patenting do not extend the patent term. And because of application publication — these files were all open throughout prosecution.

Rosenkranz then went further, attacking Judge Alsup’s finding of prejudice – a required element of laches – as unsupported by evidence. “The whole prejudice concept … is totally made up.”  Google presented no witness at trial who testified to being “caught unaware” by Sonos’s late-claiming. Sonos had presented its technology to Google in 2013 and warned Google in 2016 that Google’s new products were infringing Sonos’s  broader patents. Rosenkranz contended that Google could not plausibly claim it was ambushed by a late-arising patent claim.

Bagatell’s response was that, even if the disclosure was available, Sonos didn’t claim the overlapping-zone invention specifically until 2019. During the 2007–2019 period, “lots of other people, including Google, were making products” implementing overlapping zones. Sonos could have pursued those claims earlier but “elected not to”. By waiting 13 years to claim that feature, Sonos allowed an entire industry to develop and “the world had grown around it”. In Google’s view, the prejudice is that intervening development – companies invested in technology that Sonos only later encapsulated in a claim. This is the classic “intervening rights” type prejudice that prosecution laches is meant to prevent. Bagatell argued that Sonos had “no excuse” for taking so long; the only logical reason was that perhaps even Sonos didn’t see the overlap idea in its own spec until Google demonstrated a market for it: “I think the logical reason is they didn’t see it in their own specification.”

Judge Prost pushed back against Bagatell on this prejudice issue — assuming written description support — “the public had notice” of the potential claim. Bagatell answered that even so, prosecution laches can still apply. There are cases, he noted, where even timely-published applications and ultimately granted patents have been held unenforceable for unreasonable delay. Publication mitigates the “submarine patent” issue, but it doesn’t eliminate prejudice under laches if the patentee sat on certain claims. The patentee “still [has] to diligently prosecute” all meaningful claims, he argued. Here, Sonos “took 13 years” to present the overlapping-zone claims – a delay Bagatell characterized as far beyond the 8–9 year periods in prior laches cases. No one knew Sonos would claim that specific invention until it finally did. In Google’s view, that lengthy deferment of claiming is inherently prejudicial in an industry as fast-moving as tech.

The “Shootout” Summary Judgment and Judge Alsup’s Mid-Trial Pivot

A substantial amount of time was spent on Judge Alsup’s “shootout” procedure for summary judgment. In Alsup’s courtroom, each side was invited to file early summary judgment motions – essentially a cross-motions “shootout.” Here, Sonos had moved for summary judgment of infringement (on one patent claim), and Google in its opposition argued that there can be no infringement because [the claim] lacks written description support.   Alsup rejected the written description defense on summary judgment, but with some procedural confusion.  In particular, it isn’t clear if he (1) fully rejected the written description defense as a matter of law – treating it as a total loser in the case; or (2) merely concluded that it was not ripe for summary judgment, with potential issues that could be resolved at trial.   Prost acknowledged it was “funky… because of the way the shootout worked.”

The importance of this is a waiver/forfeiture question.  If it was merely a denial of summary judgment then Google should have raised the issue again at trial (it did not) in order to preserve the issue for appeal.  Judge Bumb astutely asked “Wasn’t it up to you to pursue that [defense] at trial?”

  • Sonos’s view: The denial of summary judgment simply meant there were factual issues for trial (the court “freed Google to try the case” on written description if it wanted).
  • Google’s view: Judge Alsup granted summary judgment in Sonos’s favor on written description, finding as a matter of law the disclosure was adequate – which removed the issue from the jury. Google contends it “accepted that ruling” and simply reserved its right to appeal. Only when Judge Alsup reconsidered did the issue re-enter the case.

Judge Prost commented, “when I look at the court’s order, it is not clear. It seems to sort of say both.”

Further, Rosenkranz emphasized that Google’s counsel never raised “new matter” as an issue until the judge did so mid-trial. In fact, when Alsup asked Google’s lawyers about it at trial, “Google said, ‘oh, Your Honor, we hadn’t thought of that ’til you raised it.’”  Thus, Rosenkranz argued, Google forfeited any new-matter theory by not pressing it, and “the rules are no different just because the judge came up with the theory on his own.” According to Sonos, Judge Alsup’s actions violated basic fairness and procedure – a judge cannot introduce a new invalidity theory sua sponte after the close of evidence, especially having earlier suggested the patent was valid.  (Google’s expert had apparently noted the specification amendment in his expert report, but did not argue that it was consequential).

This setup suggests to me a good chance of remand for further factual finding based upon disputes of material fact.

“Floodgates” and Equitable Powers: Judge Bumb voiced a policy concern: “Why [are] you so eager for us to make findings on prosecution laches? Don’t we run the risk of opening a floodgate on this issue?” If the Federal Circuit endorses such a laches theory, could it embolden many accused infringers to raise laches whenever a patentee used continuations over a decade or more? Bagatell downplayed that fear, calling this case “highly unusual.” “Every one of these prosecution laches cases is unusual,” he noted. He pointed out that thousands of patents issue after lengthy prosecutions, and Google is not arguing all such patents are suspect. Rather, in his view, Sonos’s fact pattern stands out: 13 years and a late claim of a feature learned from an accused infringer. Bagatell suggested that at some point (perhaps 8–10+ years of delay) a patentee has “some explaining to do”, but not every long prosecution is laches. Here, he argued, Sonos’s explanation was lacking, and the circumstances (a late-claimed feature and an affirmative finding of new matter) are rare.

Rosenkranz responded in rebuttal that Google is asking for a new playbook for infringers: “All you have to do to follow this roadmap is point to six years of delay and raise some fact question about written description.” Nearly any long prosecution could be attacked with hindsight arguments that the original spec didn’t expressly disclose a later-chosen claim limitation. Rosenkranz characterized the written description dispute in this case as utterly routine – “a garden variety written description issue. I’ve won written description issues like this as a matter of law” in other cases. In his view, there was nothing extraordinary about Sonos’s claim or its support; what was extraordinary was Judge Alsup’s willingness to deploy laches to nullify a jury verdict.

Before Bagatell concluded, he reiterated that none of the procedural irregularities Sonos complained of tainted the laches ruling itself. Laches was pleaded and briefed; it was “entirely kosher because it was before the court at all times”. Sonos’s real gripe was with the handling of the written description issue at trial, which “none of that relates to prosecution [laches].” Bagatell’s implication was that even if the appellate court finds fault with Alsup’s process on the invalidity issue, it could still uphold the laches determination on the (extensive) record compiled in the bench trial on laches factors.

= = =

A good amount of time in oral arguments was spent on questioning from Judge Prost on what I see as a red herring—whether there is some problem with Judge Alsup finding both a written description problem (invalidity) and laches (unenforceable). Judge Prost pressed Google’s counsel on the logical tension in these alternative holdings: if the patent claims are truly invalid for lack of written description, then “where is the prejudice” to Google from prosecution delay since “Google was not going to be liable for any money” anyway? Conversely, if the claims are actually valid (meaning the 2006-2007 disclosure did provide adequate support), then the invention was publicly disclosed and Google had notice before launching its products in 2015—so again, where’s the prejudice? While Google’s counsel acknowledged this “fundamental tension,” Judge Bumb correctly noted that trial courts routinely “rule in the alternative” to cover all bases. In my view, this line of inquiry, though intellectually interesting, distracts from the core question of whether Sonos’s strategic prosecution conduct warrants the application of laches as an equitable remedy.