Day 1, Panel 1: Mark Lemley’s Annual Patent Law Update
26th Annual Berkeley-Stanford Advanced Patent Law Institute
December 4, 2025
B-CLE Recording (CLE: FREE) | Youtube Recording | Agenda | Event Resources | Speaker Biographies | Download Panel Transcript
Download the Panel Transcript and Materials here.
Speaker
Mark Lemley, Stanford Law School
Resources
Panel Summary
Wayne Stacy opened the session by welcoming attendees and noting that the proceedings were being recorded for archival purposes. He then introduced Professor Mark Lemley, who led the audience through a rapid survey of Federal Circuit decisions from the past year. Lemley characterized the year as one of “normal science” — meaningful developments, but not a fundamental rewriting of patent law. He noted that the more disruptive changes appear to be coming from the Patent Office rather than the courts, with the Federal Circuit largely reinforcing established doctrines across patentable subject matter, definiteness, written description, obviousness, and claim construction.
Lemley highlighted written description as the most significant area of development this year, noting an unusual pattern in which the Federal Circuit reversed jury verdicts and denied JMOLs in no fewer than three cases — a remarkable occurrence given that written description is nominally a question of fact. Cases like Duke University v. Sandoz and Seagen v. Daiichi Sankyo illustrated the court’s strict approach to genus claims, requiring patentees to provide sufficient “blaze marks” to identify which members of a broad claimed genus actually work. On the defenses side, Lemley argued that the court has effectively gutted both equitable estoppel (Fraunhofer v. Sirius) and prosecution laches (Google v. Sonos), making those defenses nearly impossible to sustain in litigation going forward.
Key Learning Points:
- 101 Patentable Subject Matter: Using existing AI or computer tools as mere instruments is insufficient for patent eligibility — claimed inventions must demonstrably change how technology functions, not simply invoke it (Recentiv Analytics v. Fox; Rideshare v. Lyft).
- Written Description — Genus Claims: Broad genus claims require either a representative number of working species or a common structural feature distinguishing those that work; vague Markush groups without guidance on which members function will fail (Duke University v. Sandoz; Seagen v. Daiichi Sankyo).
- Prosecution History Estoppel: Statements made during prosecution of any patent in a family — even after a related patent has already issued — can be used to narrow claim scope retroactively across the entire family (Barrett v. Fortress Iron).
- Defenses Weakened: Both equitable estoppel and prosecution laches have been rendered nearly impossible to invoke as litigation defenses, as courts now demand a very high evidentiary burden showing specific reliance on a patentee’s silence (Fraunhofer v. Sirius; Google v. Sonos).
- Damages and General Verdicts: The Federal Circuit has signaled a strong preference for special verdicts over general verdicts in multi-patent cases — if a general verdict cannot be reliably apportioned after one patent is invalidated, the entire damages award may be thrown out and retried (Optus Cellular v. Apple).
Program Transcript
Key terms: Core Patent Law Doctrines, Patentable subject matter (§101), written description, enablement, definiteness, obviousness, claim construction, infringement, doctrine of equivalents, Procedural & Administrative, PTAB, IPR (Inter Partes Review), JMOL, Daubert, prosecution history estoppel, prosecution laches, equitable estoppel, ex parte reexam, collateral estoppel, Claim Drafting Concepts, Means-plus-function, Markush groups, Jepson format, comprising/consisting of/consisting essentially of, nonce words, lexicographer doctrine, prosecution disclaimer, Damages & Remedies, Apportionment, reasonable royalty, lump sum, attorneys’ fees, deterrence sanctions, general vs. special verdict, Genus/species claims, blaze marks, polycrystalline diamond compact, CRISPR-Cas9, pharmaceutical evergreening, domestic industry, Genus claiming limits (post-Amgen), AI patentability, ITC domestic industry, abstract idea doctrine, PTO guidance vs. Federal Circuit precedent
This panel, Mark Lemley’s Annual Patent Law Update, was the first panel of the first day for the event, 26th Annual Berkeley-Stanford Advanced Patent Law Institute. This event was hosted by Berkeley Center for Law & Technology, UC Berkeley School of Law, and Stanford Law. Mark Lemley (Stanford Law School) was the speaker for this panel.
[WAYNE STACY]
So I’m Wayne Stacy. I’ll be helping with a lot of this morning. Uh, you’ll hear with me, from me this afternoon, with Director Iancu. Uh, but our first speaker this morning as we were just talking about if you don’t know, you’re probably at the wrong conference, and that’ll be Mark Lemley.
Um, but before I bring Mark up, what I wanted to do is make sure you see we put the QR codes on the, the sides for your CLE credit. And I wanted to let you know it’s being recorded. One of the goals for the two universities is to make sure that we, you know, we retain the knowledge.
So if you’re asking questions in certain sessions, just remember that. We’ll get the mic to you so you can ask them so we can keep all of this. So with that in mind, I’m gonna get this started, and Professor Mark Lemley from Stanford.
(applause)
[MARK LEMLEY]
All right. Thanks everybody for coming.
So most of you, I think, have done this with me before. Uh, and so you know the deal, which is I’m gonna go through everything that happened in the Federal Circuit over the last year pretty fast because there’s a lot that happened in the Federal Circuit last year.
You do not need to take notes. There is a paper that has all of these cases, with the exception of, I think, one that was decided on Tuesday.
I’m basically gonna kind of follow through in order. Um so overall uh, there’s a lot of interesting stuff going on, more than in maybe in the last couple of years in the Federal Circuit, although I would say it is, for the most part, normal science, right?
Uh, there are some interesting things you need to know about, but they are basically, we are not, the Federal Circuit is not fundamentally rewriting the rules of patent law as we’ve seen the Federal Circuit or the Supreme Court do in, in previous years. This year, that appears to be the job of the Patent Office. More on that later this afternoon and tomorrow. So we got a lot of cases that fit in sort of what I think of the heartland of patent law.
There’s a lot of written description, there’s a lot of obviousness, there’s a lot of claim construction, damages, so we’ll talk about some interesting things.
[WAYNE STACY]
Let me start, as always, with patentable subject matter. And here, as it has for the last several years, things have settled down into a pretty predictable pattern. You might not like the pattern and the cases don’t always get the results right, but they have a pretty basic feel to them that we’ve seen recur for each of the last several years.
Uh, so in Rideshare versus Lyft one thing the court notes and this has, it has said several times but it is worth repeating: the patentable subject matter eligibility guidance that the PTO issued in 2019, which basically tried to sort of change the law a bit of patentable subject matter, it is not the law. The court is not bound by it, it does not follow it. So you might get a patent using that guidance, that does not mean your patent is valid under 101. The Federal Circuit is not gonna give it deference.
And here in Rideshare, the court says as it has in many cases you get a, you’ll, you’ll survive patentable subject matter if your invention changes the way the technology functions in the computer environment. But if it just uses a computer in its ordinary way, invoking the computer as a tool then it’s not patentable.
And that’s what the court said here in the in the Rideshare context. It said something similar in the United Service SS– versus PNC Bank case. Here, the invention is depositing a check using a handheld mobile device. The court said, “You know what?
That could conceivably be a new and patentable invention if you’ve got some novel way of improving the technology, but you don’t have that novel way of improving the technology in the claims.” And so here’s the second thing that we see again and again in 101 cases, right? If you point me, if you wanna point me to kind of novel technology that changes the way the thing operates, you’ve got to actually find it in the claims.
It’s not sufficient to say, “Somewhere in my specification, I came up with a novel algorithm, but I didn’t limit my patent claims to that algorithm.” Uh, also of note is Recentiv Analytics versus Fox.
Uh, this is of note primarily because it is the first AI case to make it to the Federal Circuit under 101. The patent fails under 101, but it doesn’t fail because it’s AI, and it is not an indication that AI is not going to be patentable. It fails because the inventor here doesn’t actually claim that they did anything novel with AI. They claim they used existing AI technologies, and the court says, just as using existing computers as a tool, using existing machine learning algorithms as a tool is not gonna be sufficient to get you patentable subject matter.
Outside the computer context, of course, things are somewhat different. So the notable case here is US Synthetic versus International Trade Commission. And here the claim was to a chemical a polycrystalline diamond compact, an actual material. The ITC said, “You just defined that material in broad functional terms, and so that’s an abstract idea.”
And the Federal Circuit said, “No, it’s a physical thing. It can’t be an abstract idea,” right?
Abstract ideas don’t apply to products of nature, right, or other physical products. There might be other challenges you could bring to that claim, but patentable subject matter is not gonna do it. All right.
Definiteness. You’ll recall if you’ve been here before, a couple of kinda basic things that we see recurring in definiteness. One of which is the software means plus function language cases, right?
If you write your claim in means plus function format in the software world, you’ve got to show me an algorithm in the specification as the corresponding structure. And you might think you have not written your patent in means plus function format because you didn’t use the magic words, means for doing X, but the Federal Circuit has repeatedly told us, “If you give us language that is functional, that doesn’t have structure the mere existence of so-called nonce words like mechanism or means, is not gonna take it out of means plus function.” That’s true in Fintiv versus PayPal, where the patent claim term is payment handler. The court says, “A payment handler is not a thing, it’s simply anything that handles payments.
Handler here is a nonce word telling you some way of handling payments, and so we got to treat it as means plus function. We gotta go look to the specification in order to find an algorithm.” And there we don’t find an algorithm, we find instead a description of the results of applying an algorithm. Court says, “That is not sufficient.
You’ve got to actually give me the algorithm itself.” In Sierra Wireless versus Sysvel, the court confronts the interesting problem of two limitations in a patent claim that appear to be mutually exclusive.
[MARK LEMLEY]
The PTAB here said, “Claim is indefinite because it says, ‘Do X and don’t do X.'” Uh, Federal Circuit says, “No, that’s not indefinite. It’s quite clear.
It says, ‘Do X and don’t do X.'” Um, right? That’s so we reverse the indefiniteness rejection. It’s hard to imagine there’s anything the patent claim actually covers, but you got your invalidity rejection reversed, so yay, I guess for the patent owner. And final point and the other sort of case that we see here in definiteness coming around a lot are value judgment terms.
So in Akamai versus Media Pointe the relevant uh claim language refers to the optimal or best way of doing something. Pick the optimal or best path for routing. And the court says, “I got no idea what the best path for routing is.” If you just as in other cases where you’ve told me, sort of make something attractive, there’s no way for us to judge that and therefore it’s indefinite.
That’s particularly true in this case, the court says, because there are multiple different possible answers to the question of what is best. And the patent itself gives us different and conflicting guidance. There are different things you could choose, and the patent claim simply doesn’t choose it.
All right. Written description.
This is actually probably the sort of biggest development in the year. There is a lot of activity in written description.
Uh, it is almost all invalidating patent claims, and a remarkable number of the cases are cases where the Federal Circuit reverses the refusal of a, a district court to grant a JMOL after a jury had found n– uh, no written description violation. Now, given that written description is nominally a question of fact, although no one actually really believes it is treated as a question of fact, this should be a truly extraordinary thing, but it happened no less than three times this year in the Federal Circuit. Uh, so we’ll start with Duke University versus Sandoz, one of the reversals of the, of the jury verdict and the denial of JMOL. This fits in a category of cases that we’ve seen a lot since Amgen versus Sanofi, both in enablement and in written description, which is you’ve got a whole, you’ve got a, a patent claim that covers a p– reasonably broad genus, right?
There’s a whole bunch of things. Here it’s over 4,000 compounds that might cause hair to grow.
[WAYNE STACY]
You don’t actually tell us which ones will work and because you haven’t given us either a representative number of species that do work or a structural feature that’s common to the ones that work and distinguish them from the ones that don’t, you have not described the full scope of the genus and so you can’t have a patent claim to the genus. The court here talks about, has talked, I’ve always talked about blaze marks, right? Give us kind of signposts or blaze marks along the trail that help us to find the ones that will work and the ones that don’t work. Here they take that metaphor perhaps to an extreme extent saying “The patent did not provide sufficient blaze marks to direct one to the proposed tree in particular and doesn’t teach the point at which one should leave the trail to find it.”
Brita versus International Trade Commission. Here’s a written description violation that’s about the kind of reaching beyond the things you actually did effectively describe in the claim to cover some things you didn’t. So this is the Brita water filter that you’re probably familiar with. The gravity flow filter technology, the court it describes kind of flowing through carbon to filter out impurities in the water.
But the patent claim says, covers various types of filter media including carbon blocks. And the court says, “You absolutely enabled and described carbon blocks in your specification.” Your patent lawyer got a little greedy and they said, “We don’t wanna limit ourselves to carbon blocks, maybe there are other kinds of filters, but you didn’t actually give us any information about those other kinds of filters.” And so that broader claim fails for written description because we don’t know you don’t possess other kinds of filters.
In Mondis versus LG Electronics similarly kind of a reach beyond question, here there’s a patent claim that originally says, “An identification number for said displayed unit.” There’s a rejection on prior art grounds, and they change the patent claim to say, ‘At least a type of display unit.’ And the court says, “Well, you’ve broadened the universe here, right?” So now it could cover other things besides a type of display unit. “But you haven’t told us what those other things might be,” because it wasn’t in the specification, it was changed in prosecution. And therefore, the court says, ‘Lack of written description,’ because you don’t tell us what else might be covered by the words, at least.
Case decided Tuesday, Seagen versus Daiichi Sankyo, another case in which the Federal Circuit reverses both a jury verdict finding no written description and a refusal to grant JMOL. Um, eh, and in the– in the Seagen case here again, we’ve got a large genus, the court says covers potentially at least 47 million different pep– uh, peptides.
“you haven’t actually taught us kind of which of the 47 million different peptides are the ones that would work. You’ve kind of given us sort of Markush groups that you could just draw from without telling us what’ll work within each. And therefore, as of the time that you did this, you can’t claim priority to your invention ’cause you didn’t have adequate written description.”
It does not help in this case that the named inventors admitted they didn’t actually contemplate the defendant’s product, right, which is within the scope of the claims. And in fact the patent claim was changed in the course of prosecution only after the defendant invented the thing in question, right?
So, this is one of the cases where we’ve seen people try to expand their patent claims to cover what the defendant is doing. And those are, I think, particularly likely to run into written description problems. All right.
(papers rustle)
In re Xencor is an interesting one. This is a case the f– the Patent uh, Office had originally rejected the claim as a means plus function claim on the theory that means plus function claims always fail written description because they cover the structure actually disclosed and equivalents thereof, and you didn’t describe the equivalents thereof.
That can’t possibly be right. And I and some other people wrote a brief on appeal on the first appeal saying, “That can’t be possibly right.”
Patent Office actually agreed, changed its mind, pulled it back, but it nonetheless still rejected the claims on other grounds. Uh, here, the Federal Circuit reverses uh sorry, sorry, affirms the PTAB’s finding that there’s lack of written description, and the reason there’s lack of written description here is that the preamble for the claim which is written in Jepson format talks about treating a patient with a disease, and here’s my improvement.
They actually do describe the improvement, but the court says interestingly, they don’t have any possession or written description of the preamble, the stuff that’s supposed to be in the prior art. You don’t tell us what disease you’re treating a patient for. It’s the sort of kind of general treating of a patient with something and so it’s a kind of very curious case where the court says, “It’s not the novel part of your invention that you failed to describe, it’s the stuff that supposedly everybody knew that it was in the prior art, but you gotta describe that too, even though it’s in the preamble or you lose.”
Final written description case, and finally a patentee win, In re Entresto. Here the patent claim says two drugs are administered in combination. The patentee sort of, you gotta puts them both in a, you know, here’s a pill, here’s a pill, take them both. The defendant complexes them, kind of puts them together into a single integrated form.
The court says, the patentee sues the defendant. The District Court says “Hey, you violate written description because you didn’t actually talk to us about complexing the drugs, ’cause it wasn’t actually known that you could complex the drugs at the time you filed your patent application.” Federal Circuit reverses, they say, ‘No,’ right?
What the patent says is, ‘Put these things in combination. Taking two pills is putting them in combination.’ That was all that was required to be described because it was all that was understood from a combination at the time you filed your patent application. And you’re not required to do the impossible, which is provide written description for something that doesn’t yet exist.
It’s gonna be, I think, an interesting question, right, whether you can nonetheless then sue the defendant who does this different form that was later developed. In previous cases, Chiron versus Genentech, for instance, the Federal Circuit had suggested that that might violate written description.
Entresto says, no. And I think Entresto’s right about that, right? So you shouldn’t be required to describe something that hasn’t yet been invented although it might be the case that there are other doctrines that will prevent you sort of reaching it. All right. Enablement, Agilent versus Synthego.
And here the disclosure is that I argued and lost this case on behalf of the patent owner so anything that sounds like sour grapes probably is. So here it’s a patent to a CRISPR-Cas9 system for gene editing. The court invalidates it on the basis of a piece of prior art, which is an unpublished patent application that the prior art inventor never got to work but which sort of kinda gives a laundry list of, “Here are all the possible things you could do,” right?
“we didn’t actually, we tried it, the thing, it didn’t work. We didn’t actually get anything to work.”
One of the sort of millions of things that you could do turns out to be the one that the patentee didAll right. Now I think if you were trying to sort of get a patent on that basis, and you would’ve failed all of the written description cases we just talked about.
But the court says, “Enablement for 112 purposes is different than enablement or written description for getting your own patent.” First off, you don’t have to enable someone to make and use.
It’s enough to make it. “And you told us you could make these enormous array of chemicals, one of which turns out to be the one that the patentee figures out works.
And we also presume that written documents are enabling for 112 purposes.” All right?
So the court says, in part on that basis and in part based on sort of factual findings in the PTAB, this is enabling because it told us here are things that might work even though the prior art never actually got it to work. All right. Novelty. Here I think the case I wanna focus on is Lynk Labs versus Samsung.
This is a case that’s at the Supreme Court right now on cert, although the odds of it or any patent case being granted on cert are pretty low. Supreme Court has not taken a patent case in three years. Here the question is, what’s the effective date of a prior art patent application in the PTAB in an IPR proceeding?
You may remember IPR proceedings. They’re things we used to do to challenge patents and the. And the sort of interesting legal question is this, right? We know for in litigation that the effective date of a patent application is the date you file it as prior art.
So you file it, it is prior art as of the date you file it. The IPR process allows you to challenge on the basis of patents and printed publications. Here, the prior art application is ultimately abandoned.
It is published but it never actually turns into a patent. So it’s not a patent. It is a patent application, which is not expressly listed in the list of things you can cover in IPR.
It is a printed publication, but it’s published, of course, 18 months after it’s filed. Nonetheless, the Federal Circuit says, right, “Patent applications in the PTAB oughta be treated as having the same prior art date as patent, as same patent applications in court which is as of the date you file the application, it’s prior art.”
Therefore, this is prior art, and therefore the LinkLabs patent is invalid.” I mean, I think that’s right.
It certainly is a policy matter. It seems weird to say we’re gonna treat the same prior art differently in IPR and others, but I, you can imagine the Supreme Court looking at this and saying, “Gosh, it says patents. It doesn’t say patent applications. So we’re not gonna count it.”
We’ll see. All right. Obviousness. Let’s talk about Bayer versus Mylan.
Here, interestingly, the court affirms the Board’s finding that the phrase clinically proven effective was non-limiting where you didn’t actually add any new and non-obvious functional relationship with an already-known method of treatment. So this is one of the class of pharmaceutical evergreening cases.
So I already have a drug. I have a patent on the drug.
The patent is expiring. I’d really like another patent on the same drug. So I file a patent that says that same drug I’ve already been using where it is clinically proven effective.
And the court says, “That’s actually not a new real limitation. That’s something that sort of could create to make your new invention non-obvious. All you’ve done is sort of discover or claim something that was already part of that product.” You already had the chemical.
You were already using it. Adding that claim seems like really an inherency problem.
The court notes the evergreening problem, right? “we gotta prevent the indefinite patenting of known products and methods,” it says, “by the simple inclusion of novel yet functionally unrelated limitations.
And we don’t want you to claw back from the public domain an anticipated method of treatment merely by adding a limitation that the method subsequently performed well in a clinical trial.” And that seems kind of right as a policy matter. All right. Ankura versus Roku.
This involves secondary considerations and the importance of the nexus evidence. So you may recall if I wanna prove secondary considerations like commercial success, I’ve actually got to show that the patent was the reason for the commercial success.
In Ankura, the court says that nexus requirement is much weaker, much more attenuated where the evidence is not commercial success but licensing. Because it says, “Look, the reason people license patents is because of the patented features and the inventive features, so we can presume that there’s a nexus between the sort of patented features, the novel features of the invention and the reason you bought it.”
They do note the sort of fact that sometimes people have been known to take licenses and patents in litigation like this one was to avoid nuisance value settlements. And here what they say is, “We don’t think that is this case because the settlement value in the prior license far exceeds the cost of litigation even though the defendants were aware of the prior art.” So the previous defendants in the case who had the same arguments and could have made it obvious were willing to pay a lot of money to license this. That seems like evidence that it is non-obvious.
At least it was non-obvious to that particular defendant. All right. Claim construction. Lots going on in the world of claim construction.
I’ll start with Aortic Innovations versus Edwards Life Sciences. This is one of several cases where we see the patentee treated as their own lexicographer.” Here the relevant claim language is outer frame and the Arctic patents define outer frame to include both serial and dual frame embodiments.
But they consistently refer to the outer frame as a self-expanding outer frame. And the court says, “By consistently using the term self-expanding to describe and define a outer frame, you have acted as your own lexicographer. People would understand this to mean only self-expanding frames, and so that’s what your claim is limited to.” In Barrett versus Fortress Iron, the patent claim term in question is non-integral bosses.
I think we’ve all probably had some non-integral bosses in our lives, but these non-integral bosses involve sort of fastener attachments, and the question is is the patent claim limited to fastenerless bosses? Here the court says, “You haven’t clearly and unmistakably disclaimed bosses with fasteners in the patent specification itself. But in the course of patent prosecution, you ultimately do affirmatively disclaim them.”
What’s notable about this is that the prosecution in question is not the prosecution of this patent. It is the prosecution of another patent in the same family, and it is the prosecution of another patent in the same family that happens after this patent is already issued. So the patent is already issued with the term non-integral bosses.
Later, in another family member, you say, “Oh, no, that doesn’t include ones without fasteners.” And the court says, “We can apply that retroactively to this patent too,” ’cause they’re all part of the same family, and so they’re part of the kinda same prosecution history. So important kind of prosecution safety tip, right? Anything you say can and will be used against you, not only in this case, but in your own prior cases that are part of the same family.
All right. Fmc versus Sharda.
Here the sort of question is is the term composition limited to stable compositions in an insecticide patent? There’s a bunch of disclosures about the importance of physical stability in the provisional application but the patentee takes them out when they file the actual application.
And the court says, “The fact that you changed from the provisional to the final is highly significant in interpreting the terms. It suggests that it doesn’t have to be limited to stable compositions. You affirmatively changed your patent to broaden it.
You probably lost priority in the course of doing that to your provisional application. It’s a new thing, and you’re adding you’re broadening the scope of it, but as of the date that you file your new application, we’re not gonna bind you to things you once said in the provisional, but have affirmatively walked away from later.
Uh in LabCorp versus Qiagen the uh, here is just a sort of useful reminder that the Federal Circuit tells us time and again, and as district courts seem to keep forgetting if there is a dispute about patent claim uh, meaning, it is a dispute for the court, and the court needs to resolve it even if it happens after the Markman hearing. You can’t send it to the jury.
In this case, the relevant claim term was identical. Uh, the Federal Circuit says, “The jury can figure out what identical means and allows the patentee to argue that identical might mean identical to a portion rather than the whole thing.” Uh, Federal Circuit says, “Identical means the same. It doesn’t mean identical to a portion.
Uh, that claim construction is wrong, but more to the point, the jury shouldn’t have been allowed to hear claim construction, because that’s not something juries are supposed to do. All right.
Um, Eyetherapies versus Slayback. Um, this is uh, the Federal Circuit says, “This is another kind of lexicographer change your meaning.” Here you changed the meaning of the phrase consisting essentially of.
This causes me personal pain, because I feel like there are only three claim terms in all of patent law that have settled established meanings, right? Comprising, consisting of, consisting essentially of. And now you said, “Well, sometimes when you say consisting essentially of, you don’t actually mean what we always understand consisting essentially of to mean.” The facts are a pretty good case for that, right?
So here I was supposed to exclude a particular chemical ingredient in addition to the one I was claiming. Patentee, the Patent Office said, “Hey, prior art had this one. I want to exclude the other ingredient,” so the examiner and the applicant agree that we can do that by changing the claim preamble to consisting essentially of, which should get rid of non-essential ingredients. Patentee then turns around and says, “Oh, I should be able to cover this additional chemical because it’s not essential.”
And the court says, “No, you’ve essentially disclaimed any effort to reach the very chemicals that you got rid of before. All right. Um Alnylam versus Moderna.
This is the COVID vaccine patent litigation. Here’s another lexicographer case. So the patentee says, “I didn’t intend to limit branched alkyls to particular embodiments.” It turns out if you wanna argue that you are not acting your own lex– as your own lexicographer, you should not in the patent write a section called Definitions, highlight the term uh, and then set it off in quotation marks with a specific quoted definition of the term uh, which is exactly what happened here.
[MARK LEMLEY]
Not surprisingly, the Federal Circuit found that you had here quite expressly acted as your own lexicographer. All right. Um, uh, versus Abiomed, just note the Federal Circuit says if I have– If I actually engage in prosecution disclaimer with respect to one claim term, doesn’t necessarily mean I have disclaimed the same con– uh, with respect to other patent claim terms if those patent claims are different. That seems right to me here.
All right. Um, Kids2 versus TOMY.
So this is a patent on a kid’s bath for like small infants and very small children. Uh, and so there are two patent claim construction issues that divide the Federal Circuit. One is distal edges joined at a bottom surface apex.
So got edges and they are joined at a bottom surface apex. Federal Circuit says, “Eh, joined, it doesn’t mean they actually have to touch each other,” right? There could be something in the middle. So in this case, it’s a distal edges, and there’s something in the middle, and the court says the something in the middle is the thing that’s joining them, and so they are joined.
The second issue is a raised portion of the bottom surface that prevents a sliding child from sliding down, whether that constitutes a seating surface. There’s basically a little kind of lip that prevents the kid from sliding down into the water and drowning, but it’s about that big. Uh, and the Fed– Federal Circuit says, “Yeah, that could be a seating surface.”
Uh, Judge Chen in dissent says, “No, a seating surface is something you sit on.” It’s not something that’s sort of like, kind of like a speed bump in the way of kind of sliding down. But that’s the dissent.
Also, in this case, the seating structure has to be two seating surfaces disposed at differing inclinations. Uh, here, the Federal Circuit says the flat bottom of the tub can be one of those inclinations.
Judge Chen, in dissent, says, “No, inclinations are things that kind of like go up,” right? There ought to be kind of two different angles of up.
Zero is not an inclination. You could go either way on that, I suppose.
All right. Uh, infringement, um, Regeneron Pharmaceuticals versus Mylan.
[WAYNE STACY]
Here, the court says if you wrote two different patent claim elements, right? You sort of listed them out.
We will presume that the infringing product must include two different things as well. I’m not sure that presumption’s actually justified anywhere in prior Federal Circuit case law, right? And so defendant here does one thing that performs both of those functions. And the question is, right, can that satisfy both of those claim elements?
Court presumes that the answer is no. Now here, it turns out there’s some evidence in the kind of specification that suggests that we really were thinking about two different things. Uh, but I’m not sure the court is right to say there’s a, there’s a presumption here.
All right. Doctrine of Equivalence Colibri versus Medtronic involves prosecution history estoppel. And here, the useful thing to note, which I think is consistent with what the court has said in the past but it is worth remembering prosecution history estoppel applies if I narrow my patent claims, all right?
One way the court says you can narrow your patent claims is to cancel a broader independent claim, all right? So that has always been true I have an independent claim and then I have a narrower dependent claim.
I cancel the independent claim, and I promote the new dependent claim. It is narrower.
I didn’t actually physically narrow that claim, but I narrowed the sort of broader scope. In Colibri, the court says that can also be true of two different independent claims. So if you have two possible ways here of pushing a valve out from a syringe and one is claimed in one way, the other is claimed in another way, and you cancel a claim to one of those two ways, you can’t use your patent claim to the other one to try to get back the very thing you canceled, ’cause the very principle of prosecution history estoppel is if you wanted a claim to that one, you should’ve actually asked for it, right, and not given up when the Patent Office challenged it.
All right. Uh, International Trade Commission Lashify versus ITC I think cements the idea that when we talk about domestic industry practicing the patented invention, we don’t really mean it. Um, so we changed the law like 30 some years ago to give up on the idea that we actually manufacture things in the United States, and now it’s sufficient that you have a, a patent licensing business in the United States.
Uh, here in Lashify, the court says the domestic industry doesn’t actually have to include people actually working on the patented invention itself. It’s enough to include a domestic industry of sales, marketing, warehousing, distribution, quality control.
Your accountants are a domestic industry. Your lawyers are, if they work in-house, a domestic industry. Uh, anything that involves kind of some expenditure of money in the United States is gonna be sufficient to satisfy the, the domestic industry requirement.
All right, defenses. So here, we actually see some really significant changes.
And we’re gonna hear a little more about these kind of later on in a panel today. But I think the court has effectively eviscerated both equitable estoppel and prosecution laches this year. In equitable estoppel in the Fraunhofer-Gesellschaft versus Sirius case the patentee knows about the defendant, does nothing for five years. Court agrees that’s a that is silence despite clear knowledge of infringement that rises to the level of effectively misleading conduct.
All right. We know each other.
We’ve talked. I know you’re infringing. I haven’t done anything.
But the court says that’s insufficient for an equitable estoppel because you, defendant, haven’t proven that you– know you engaged in kind of investment in this technology during that five-year period. You haven’t proven that you made that investment in specific reliance on silence from the patent owner rather than for other business reasons. I don’t know how you prove that, honestly.
I mean, I guess maybe you could just sort of have lawyers always try to paper it, right? But you know, the reality of the situation is, right, I know there’s a patent. I know the patentee hasn’t sued me over a peri– a period of years. I decide to invest in a product that I would have thought was sort of what equitable estoppel looked like.
But now I have to prove you made a decision to invest in the product solely because the patentee had chosen to be silent in choosing not to sue you. That seems like a really hard burden to meAlso in the really hard burden to meet is prosecution laches.
In Google versus Sonos the the District Court finds that Sonos was guilty of unreasonable and in-inexcusable delay in the prosecution of their patents. Um, there is in fact a substantial period of delay, the Federal Circuit agrees with that. But the Federal Circuit says, “Yeah, but the patent application was published. And if the patent application was published, you should have known about it.
And if you should have known about it, you couldn’t be prejudiced by the fact that it might turn out 12 years later that they would continue in prosecution and get this patent. And the practical effect of that, I think, is to make prosecution laches almost impossible to apply as a litigation defense, except in the extremely rare cases where the patentee like withholds from publication under the kind of rules that allow you to do that, right? If the patent application is published, I think you’ve now got to assume, even decades later, maybe a patent will come out, and there’s nothing I can do about it. That doesn’t mean there’s no prosecution laches, but it probably means it’s going to apply essentially only if the patent office tries to apply it in-house, not as a litigation defense.
One place where they did that is Hyatt versus Stewart. Gil Hyatt is still trying to get patents through from his 1995 pre-GAT applications ’cause then he’ll get 17 years of protection. The patent office here throws out a bunch of applications there that are still pending after 30 years, and the Federal Circuit affirms. All right, remedies.
So here, interesting stuff going on in attorneys’ fees. A lot of grants of attorneys’ fees, including some decisions reversing a refusal to grant attorneys’ fees in the District Court, which is pretty rare.
EscapeX IP versus Google is a great laundry list, here’s your first ethics for the day, what not to do if you’re a litigator. EscapeX sues YouTube Music as an infringing product.
It turns out that YouTube Music doesn’t do the thing and they wanted to sue YouTube Video instead, so they ultimately have to sort of like drop it and sue YouTube Video. Then it turns out that YouTube Video’s actually been doing this since before the patent was ever issued, which is a bit of a problem for your validity theory. Then it turns out that your patent has already been held invalid under 101 in a different court in a different litigation.
So all of that seems maybe a reason to abandon and run away. So to do that, EscapeX files a ‘joint stipulation of dismissal,’ where each party pays their own attorneys’ fees, but they forgot to ask Google to sign the joint stipulation of dismissal, and Google did not in fact agree, that it would bear its own fees and costs. So when Google files a motion, they then double down, filing a Rule 59 Motion for Reconsideration of the fees order, coming in to try to sort of argue that really all of this was effective.
On appeal, that fails. On appeal, they actually come up with what seems like, if it were factually true, might be a plausible argument on the merits. But it turns out they didn’t raise that argument at any time in the District Court, and so the Federal Circuit says it’s waived. And the Court also notes that they didn’t bother to respond to Google’s efforts to meet and confer that might have resolved this.
Court is very clearly happy to sort of grant fees and to grant fees on the reconsideration of fees motion, which it said was frivolous, and you should have just kind of cut your losses uh, back then. In FutureLink versus Realtek, here the Court reverses the denial of attorneys’ fees.
This is an interesting case which has got some antitrust implications, in which turns out in the course of litigation that FutureLink had entered into a contract with a third party, MediaLink, in which MediaLink agreed to pay it a bunch of money if it would go sue the defendant. And that’s potentially problematic. The District Court sanctions them for failing to have told us of this by turning a voluntary dismissal without prejudice into a voluntary dismissal with prejudice.
The Court, Federal Circuit says that makes the defendant a prevailing party. A voluntary dismissal with prejudice, even if it’s kind of now involuntary, means the defendant is a prevailing party, they can’t be sued again, and so they are entitled to their fees.
The PS Products versus Panther case also involves not only the award of attorneys’ fees but deterrence sanctions, that is sanctions of more than the defendant actually paid in attorneys’ fees under the inherent power of the Court. Here, there’s utterly baseless litigation. They filed 12 different lawsuits.
There’s a bunch, again, of bad facts. But the Court says you’re not actually limited to this section 285, recover attorneys’ fees; you can actually get additional sanctions, punitive sanctions, as a form of deterrence. All right. Damages, Rex Medical versus Intuitive Surgical.
Patentee wins a case and the jury gets $10 million. District Court cuts that to $1 which is rather less than $10 million, right? Because it throws out the plaintiff’s expert pre-trial under Daubert because the plaintiff’s expert failed to do apportionment. Nobody puts on any evidence of damages at all in the course of the case.
In closing, plaintiff says, “We’d like $10 million.” Jury says, ‘Sure, here’s $10 million.’ District Court says, ‘No, there’s no evidence I’m reducing this to $1.’ So, interestingly so the Federal Circuit, the patentee on appeal says, that’s not just a remittder, that’s really a JMALL, no damages JMALL.” And the Federal Circuit says, ‘Yeah, we agree.’ And it was right to do that, because you don’t actually have any evidence of patent damages. Now this is interesting, I think, because it is a key shift in Federal Circuit law.
The Federal Circuit has traditionally said, ‘There’s always some amount of money to compensate a patent owner.’ But here they are willing to say, ‘You know what? If you really don’t have any evidence, a dollar is enough.’ I will also notice as a practice matter patentee makes the wrong argument here. If it is a remittder under remedies law, you get to choose, ‘Do I want to take the $1 or do I want a retrial on damages?’ By arguing that, ‘Hey, you’ve effectively ruled against me on JMOL,’ I think they give up their possibility to a retrial, and indeed they don’t get one in response. The other sort of big thing in damages is that the big thing that didn’t happen in the Federal Circuit en banc decision in EcoFactor versus Google, which everybody expected was actually gonna do something of significance.
Turns out to do, frankly, not very much. It reverses a, I think, pretty clearly wrong decision from the panel and holds, I think correctly, that the fact that the patentee says “I believe that this small lump sum payment really constitutes 25% of royalty, ongoing royalty rate in a prior license.” Where the defendant himself says, “I’m not agreeing to a royalty rate. I don’t agree that it is, represents a 25% royalty rate.”
That doesn’t count as a royalty rate. It counts as a lump sum payment. Probably the bigger implication of EcoFactor, if we have one, is the court seems to take seriously the 2023 amendments to the Federal Rules of Evidence that strengthen the role of the judge in Daubert hearing.
So the courts, I think, are suggesting this is not a, ‘Let it go to the jury and let them weigh the credibility question.’ Federal judges need to be more involved in Daubert in deciding the merits of the kind of expert’s theory. And I think that we’re gonna see that happen and play out.
And we see that actually play out in some of the other cases since. All right. I do wanna flag Optus Cellular versus Apple and a couple of other cases.
There are now, in this year, there are three cases where, for various reasons, the district court, the jury trial in multiple patents part of that jury trial gets overturned or vacated. And then the Federal Circuit says, “We gotta throw out the whole thing because we don’t know, since you used a general verdict, right, which was you know, what the, what the damages base was dependent on.” This too is a really significant change in how the Federal Circuit has done these things.
Federal Circuit has always gone out of its way to say, “We’re gonna try to find any way to sustain the jury verdict.” And so, if there were three patents and one’s invalidated, eh, there’s still two patents and they cover the same defendant’s product.
Maybe we can afford this affirm the damages award. Now we’re doing the opposite. Three separate cases this year where the court says, “If it’s a general verdict and we can’t tell which part of it was at issue we’re gonna have to throw out the whole thing and do it again.”
And I think that’s a strong push towards special verdicts and not general verdicts. All right.
I am running out of time, but I do wanna flag one or two additional things. Uh, so MGEIMA versus Philips, another one of those cases. The sort of final thing I wanna flag, I think, involves a series of issues.
I’m mostly not gonna talk about PTAB prosecutions ’cause we’re gonna hear a lot about that later. But I wanna flag a series of issues around estoppel.
Right? So you’ve got Croy versus Groupon.
Uh, Croy Holdings versus Groupon where the court says well the decision of the PTAB uh, is not given collateral estoppel effect the, in the court because the standards of proof are different. All right?
The fact that there’s a decision of a PTAB holding a patent invalid is made under a preponderance of the evidence. That doesn’t tell us whether in court the same issue in a different patent would be invalidated under clear and convincing evidence. There’s a dissent from denial of rehearing en banc by Judges Dyke and Hughes. Um, the Ingenico versus IO Engine involves the sort of scope of estoppel for IPRs.
And here the court says “The only things that are estopped, that you can’t bring in litigation, are the things you could have brought in IPRs.” That’s what the statute says.
You could have brought only patents and printed publications. You could not have brought on sale or public use or kind of non-publication prior art. And so you are not estopped from making those arguments in court, even if you lost in the IPR, and even if it’s really parallel art.
It’s really the same thing. So, if you’ve got a publication and a sale, right, you might in fact get two bites at the apple, which I guess is probably good since you weren’t gonna win your bite at the apple in the PTAB these days.
And finally there In re Juster Technologies decided on Monday. Also potentially important as we think about what happens in a post-IPR world. The Federal Circuit says, “IPR estoppel doesn’t bar ex parte reexams.” So even if you lost in an IPR, even if you lost in litigation, you can bring an ex parte reexam.
There’s just not an estoppel effect there ’cause the statute doesn’t provide for one. All right.
With that, I will stop. And I look forward to the rest of the conference.
(applause)