by Dennis Crouch
The decision in Cheetah Omni is an important reminder regarding implied licenses. The holding solidifies the presumption that a license to a particular patent impliedly includes a license to all parents and continuations that “disclose the same inventions as the licensed patent.” This extends the court’s prior decision in General Protecht Group Inc. v. Leviton Manufacturing Co., 651 F.3d 1355 (Fed. Cir. 2011) and builds upon the parallel doctrine of legal estoppel discussed in AMP Inc. v. US, 389 F.2d 448 (Ct. Cl. 1968) and TransCore, LP v. Elec. Transaction Consultants Corp., 563 F.3d 1271 (Fed. Cir. 2009).
Cheetah Omni LLC v. AT&T Services, Inc. & Ciena Corp. (Fed. Cir. 2020), began when Cheetah sued AT&T for infringing its U.S. Patent 7,522,836 (optical communication networks).
In a prior lawsuit, Cheetah had asserted the ‘714 patent against Ciena and Fujitsu (See family tree above). That lawsuit settled with a license agreement and covenant not to sue — relevant here because Ciena and Fujitsu supply AT&T with the accused infringing devices. The question on appeal: whether the license agreement includes an implied license to the ‘836 patent.
The Ciena/Fujitsu lawsuit centered on the ‘714 patent. The resulting license agreement was directed toward the patent-in suit and also expressly called out certain family patents that were also included in the license:
all parents, provisionals, substitutes, renewals, continuations, continuations-in-part, divisionals, foreign counterparts, reissues, oppositions, continued examinations, reexaminations, and extensions of the Patents-in-Suit … whether filed before, on or after the [License] Effective Date.”
Ciena License Agreement (emphasis added). Although the list of family-members is extensive, it does not include continuations-of-parents or the specific relationship here: a continuation of a continuation of a parent-in-part (or nephew).
The Federal Circuit asks the question as follows:
In personal terms, because the uncle and grandparent of the ’836 patent, are licensed, is the ’836 patent also licensed?
I’ll pause here to note how the Federal Circuit’s question shifts the focus from the way I presented the case above. Although the license agreement discusses everything in relation to the patent-in-suit (the uncle ‘714 patent), the Federal Circuit concludes that an important relationship is the “grandparent” relationship between the ‘836 patent and the ”925 patent, neither of which were the patent in-suit. That shift to the ‘925 patent is important for two reasons (1) we have prior precedent implying a license to grandchildren patents but not to ‘uncles.’ General Protecht; and (2) the linkage between the ‘836 and the uncle ‘714 patent is tenuous because of the continuation-in-part status, whereas the grandchild lineage is straight continuation all the way down. The conclusion here is easy once the focus shifts to the grandparent-grandchild setup.
In General Protecht, the Federal Circuit found that a license to particular patent impliedly licensed the grandchild as well (continuation-of-a-continuation) – “because those continuations disclose the same inventions as the licensed patent.” On appeal, the Federal Circuit ruled that the General Protecht reasoning worked here as well:
Applying the presumption established in General Protecht provides a simple and clear resolution in this case. Because the ’925 patent is an expressly licensed patent in the licenses, the licenses also include an implied license to a continuation of its continuation, the ’836 patent. . . . If Cheetah did not intend its license “to extend to claims presented in continuation patents, it had an obligation to make that clear.” General Protecht.
Slip Op. License agreements are contracts typically interpreted under state-law (or foreign law). This particular style implied license appears to fall under federal law in finding that a patentee is legally estopped from claiming no-implied-license. “Legal estoppel prevents licensors from derogating or detracting from definable license rights granted to licensees for valuable consideration.”
Cheetah makes a couple of extra arguments here.
- The ‘836 patent covers a different/narrower invention — Doesn’t matter since “the same inventive subject matter was disclosed in the expressly licensed patents.”
- The parties all knew about the ‘836 patent but didn’t include it — Doesn’t matter because presumption goes against patentee. Further the license included a list of excluded patents and the ‘836 was not on the list.
I’ll note the 2018 decision in Cascades AV LLC v. Evertz Microsystems Ltd., 335 F. Supp. 3d 1088, 1097 (N.D. Ill. 2018) where the district court refused to find an implied license for a divisional. The court noted differences in the claims — however the Federal Circuit here holds (again) that the focus should be on the disclosure. That case was settled though so we don’t know how the Federal Circuit would react.
First AI-created Drug Enters Human Clinical Trial
On average, drug discovery and development of a single pharmaceutical takes 4.5 to 5 years, with upward of 10 years considered common. But British start-up Exscientia and Japanese pharmaceutical firm Sumitomo Dainippon Pharma have turned this on its head by leveraging artificial intelligence (AI) to reduce the timeline to less than one year—and that AI-created drug is now entering a Phase I human clinical trial.
DSP-1181, a long-acting, potent serotonin 5-HT1A receptor agonist, is intended for the treatment of obsessive-compulsive disorder (OCD). In Japan, approximately 1 million people suffer from OCD, while the disorder affects 3 million persons in the United States.
More here: link to laboratoryequipment.com
It doesn’t say anything about patents
It would be interesting to see whether Exscientia or Japanese pharmaceutical firm Sumitomo Dainippon Pharma have made any filings and exactly whom they have listed as inventors.
Quite contrary to Greg’s notion below, a human person as lab tech that merely “follows orders” and does lab work verifying a directive is NOT an inventor (legally speaking).
Attorneys processing any related applications (as well as purported inventors) should be aware that false statements as to being an inventor when one is not IS ‘fraud on the Office.’
From the article:
“Since the Centaur Chemist debuted in 2012, Exscientia has made pharmaceutical collaborations part of their modus operandi. In addition to Sumitomo Dainippon, Exscientia has ongoing collaborations with:
1.Bayer for the drug discovery in cardiovascular and oncology disease
2.Celgene for the discovery of small molecule therapeutic drug candidates in oncology and autoimmunity
3.Sanofi for the design of bispecific small molecules for metabolic disease
4.Sunovion for the discovery/optimization of novel medicines for psychiatric disorders
5.GSK for the discovery of novel and selective small molecules for up to 10 disease-related targets
6.Roche for the design of pre-clinical drug candidates in oncology.”
How many of these need be designated ‘obvious’ given that a prior art system is being used in each?
FYI, a quick published application search for “Exscientia” reveals zero hits, while for “Sumitome Dainippon” shows 121 hits.
OT, but one thing that I think is interesting is the interplay between AI and new molecule discovery.
It seems that AI is now playing a role in figuring out what new molecules to try. Imagine a world where all the AI was secret, which is where Alice brings us.
Another thought. Be very interesting to see a compilation of all the comments about patents from the show Shark Tank or Dragon’s Den, etc. Almost all the comments say that it is not worth investing in something without a patent. Mark Cuban does rail against what he calls junk patents in information processing, but he doesn’t seem to really understand the space very well. I know his background.
Be interesting too to see a correlation between the investments in the different countries that have their version of Dragon’s Den and the strength of their patenting system.
Seems like this is one area where the rubber hits and road.
Anyway, just two places where maybe the rubber meets the road.
AI and molecule discovery.
Dragon’s Den and patents. I have been surprised by how many times the Sharks or Dragons have said no patent, no investment. And this is their money and they have vast experience. And it is interesting too that you get Cuban’s comments about junk patents, but I think the thing is that it is perhaps self interested in that maybe he is like SV and just wants to take whatever he wants. I notice too that he seems interested in patents when he wants to invest and rails against patents at other times.
He is also the only Shark/Dragon that I have heard rail against patents.
The other interesting thing is that the Dragon’s Den show are all over the world with the same premise.
The “AI as devisor” angle has been touched upon (albeit not too vigorously) on the other blog.
The discussion seems stuck at the “just using a machine” level, and does not advance into the facet of a non-human devisor.
As I recall, Greg Aharonian had long ago railed that AI would de facto make all such molecular innovations obvious. If I recall correctly, his thinking is that: after all, once the “devisor” part leaves the actual human sphere of operation, any and all output must be considered nothing more than the work of the legal fiction. Or another path to the same destination is that the legal fiction now must include the AI, and since no human is actually devising the end results, those end results cannot be considered inventive.
Mind you, I do not agree with Greg’s line of thought. I do note that such thinking bears an eerie parallel to other “Big Box of Protons, Neutrons, and Electrons” thinking that certain things are not patent eligible.
“Does not advance” ==> “does not advance too deeply” (there is a push to advance so, but the conversation is not picked up).
And how do the foul judges at the CAFC come to grips with the fact that what they hold as being abstract is actually discovering new drugs that may save millions of lives.
As far as I know, in the US Sovereign, the issue of AI and obviousness (for example, whether the legal fictional person of PHOSITA absorbs fully the non-real person of AI) is yet to be a matter of first impression.
Separately, the US Sovereign HAS decided the issue of whether or not any non-real person may be the originating inventor (and person of first vesting of the inchoate right turned full legal right of a granted patent). See Stanford v. Roche.
My main point, anon, is that as the AI rivals people’s intelligence or ability to solve problems the fiction that the claims are abstract becomes more and more absurd.
And the other point I am making is what will it be like if more of this information is hidden and brought into trade secrets as is happening now.
The wa nk ies on the CAFC have no clue what reality is or what is happening in the real world. They need removal. If Trump wins big in November, it has to be a real goal to dissolve the CAFC and reform it with qualified judges and make some type of rule that would invalidate all their opinions since like 2010.
If Trump wins big in November, it has to be a real goal to dissolve the CAFC and reform it with qualified judges…
I doubt that Pres. Trump even knows that the CAFC exists. It is not a court in which he has been sued, and it is not discussed on Fox & Friends, or Hannity. If you are hoping that Pres. Trump will take an interest in the CAFC, I think that you are going to be disappointed.
I find myself in agreement with Greg on this point. The current circus is simply disassociated with the current mess of the ‘Simian in a cage, fire-hosed trained by the Supreme Court’ CAFC.
The ‘game of Trump’ is not related to the mess of IP law. And likely won’t be.
The root cause is — and remains — the Supreme Court. Any truly lasting solution then needs to incorporate both the Congress exercising its Constitutional power of jurisdiction stripping of the non-original jurisdiction of patent cases from the Supreme Court, as well as resetting a proper Article III court (in order to preserve Marbury). You are absolutely correct in that the current CAFC is simply FAR too tainted.
I hear you — without disagreement.
The AI angle does have multiple touch points, from inventorship, to obviousness, to eligibility (and the current jurisprudence in regards to eligibility is indeed in shambles, even before any AI impact registers).
I also hear you on the Trade Secrets front. In fact, more than once recently (for example: link to patentlyo.com ), I have drawn attention to the fact that the AIA sanctions a ‘higher level of power’ to the different path of Trade Secrets, and that those who seem insistent on continuing a confused and conflated view of time of protection under these two different protections, seem oblivious to the far larger elephant in the room of the new Sui Generis “right” that is in truth a submarine Anti-right. Quite evidently, those that are oh so typically against strong patent rights (like Paul and Greg, among “I Use My Real Name” folks) turn to deafness on this ‘gift’ to those that captured Congress in the writing of the AIA.
The trade secrets enable capture of employees.
+1
I have a question.
Of all the commentators, who has the most patents? So far, I win.
S. Morse asked, “Of all the commentators, who has the most patents? So far, I win.”
I am the named inventor or co-inventor on 22 U.S. Patents. All of them are utility patents.
Some are my own patents, some are for the work I did at Atari Coinop.
Some of the patents are so old that they have expired due to old age.
Some are new enough that I am certain to expire first.
NOiP,
I would not put much credence in what the poster with the shifting historical pseudonyms would say.
If you ever notice the actual content of that person’s posts, there is so very little signal and so much noise that the only takeaway possible is that our pal Shifty is only playing games.
Also, I extend congrats to you (and Shifty) for being named inventors. That being said, whether or not named inventors can intelligently add to the legal conversation (unfortunately) has no correlation. While certainly there are astute inventors who recognize their limits of intelligence (and such limits exclude patent law), and certainly there are those that do earnestly learn about patent law, there are many inventors who only think that they know about patent law (when they really do not). Sometimes being smart enough to innovate in one field becomes a curse of thinking that one knows everything about everything.
If you ever want to take a look at some humbling data, look on the USPTO website for the patent bar passage rate, and remember that this passage rate largely applies to those who have EACH successfully obtained engineering degrees, law degrees and passed state bar exams.
Notice as well here, NOiP, that the context of “has” in Shifty’s question is more than a bit ambiguous.
Shifty may well NOT be referring to being the named inventor on the most patents.
Shifty may well be referring to “has” as in “having processed” the most patents (if his line of work is in the publication department at the USPTO).
Similarly, Shifty’s context of “has,” may be the most patents tabulated in his prior art searches (if his line of work is being an Examiner at the USPTO). Note that such a tabulation need not even include reading and understanding the volume of patents included within the word “has.”
Note finally the anti-patent rhetoric from Shifty, and take that as an indicator of the line of work that Shifty may actually be in. If he is truly claiming (and not just insinuating) that he is a ‘most prolific’ inventor, would he really be so maniacally focused on being anti-patent?
I agree that your points are OT, but they are more interesting that the actual topic, so I hope that no one will mind my responding and dragging the discussion further off topic.
(1) I agree that AI for drug discovery is a poster child (one among many) for why Alice will create sub-optimal policy outcomes.
(2) Of course, in a world in which the AI tells the scientists what drugs to investigate, and then the bench scientists follow-up and test the drugs, the obviousness analysis gets more complicated. At least on the near horizon, I expect that pharmacology will still be an “unpredictable” art (although if the AI gets to the point where pharmacological outcomes are as predictable as, e.g., the flight paths of artillery shells, that will really change the obviousness analysis). Under current standards, one cannot list the AI as an inventor, so I am not sure whether the AI’s contribution to the discovery can be excluded from the prior art (it is the contribution of “another,” but not the sort of “another” who can be added to the list of inventors).
(3) The right way to solve that problem noted in #2 is to change the statute to allow the listing of corporate entities not only as the “applicant” but also as the “inventor.” In other words, in the same way that the corporate entity is the “author” of a work-for-hire in copyright, the corporate entity should be the “inventor” of a work-for-hire in patent.
Greg’s point (2) vis a vis lack of inventorship for AI does NOT mean that anything BY AI falls outside of the State of the Art consideration for use in obviousness.
His attempt at distinguishing “another” as “not the sort” is a red herring.
I also reject his ‘answer’ of (3), as such flies completely opposite the rationale in the controlling law of Stanford v. Roche.
Who knows what Greg is really thinking with his (3), as it APPEARS that he is merely mouthing the Big Pharma desire that aligns with Corporatacracy supplanting real people (and it is HIS view that is gobbledygook ‘words on words’ that simply ignores the state of the and the underlying understanding of that state). Additionally, there is NO NEED for his supplanting of real person as inventor, as the current mechanisms of assignment fully cover any legal need. By why let that fact get in the way?
Be interesting too to see a correlation between the investments in the different countries that have their version of Dragon’s Den and the strength of their patenting system.
Why would that be interesting? The investors in the (e.g.) German version of Shark Tank do not care overmuch about whether you have a German patent. They care about whether you have a U.S. patent, same as the investors here.
The German market is not large enough—by itself—to matter. Same for the Japanese market, same for the French market, same for the Canadian market, etc.
The two patents that each cover a single market worth caring are the U.S. and the PRC patents. Those are the patents about which the VC investors care, regardless of the investors’ nationality.
… and the hypocrisy of Greg leaks out a bit with his acknowledgement of localized effects of patents (against his otherwise ‘One World Order innovation anywhere is the only important consideration’ views).
Actually, Greg, that isn’t true. The dragons or sharks in the other countries like the UK do care about UK patents as the markets and products they are going after often are more modest than billion dollar home runs.
Overwhelmingly, the shark/dragon shows illustrate that patents are vital to innovation at least that is what the investors believe.
If 2 of 3 different AIs come up with the same molecules to try, then AIs are the new PHOSITA (Programs Having Ordinary Skill in the Art).
Can a company get a patent on something invented by an AI? Can they name the person who originally programmed the AI as the inventor on the patent or do they have to fess up?
Rather than a flash of genius it is flash memory.
The conflation of Programs and Persons is an interesting twist on the old fallacy that computing machines are nothing more than the proxy of real people, and for that reason (alone) cannot be the subject of patent protection — along a twisted and convoluted mix of 103 and 101 rationales.
The opinion concludes “And because we have concluded that AT&T’s products are licensed under the license agreement, we need not consider the scope of the covenant not to sue.” How do you think the court would have ruled if there was only a CNS with no license?
Be careful how you contract.
Agreed. That language is just odd. I’m trying to figure out if it was just an awkward way of trying to make the license encompass everything related up/down (which is fairly common), or if the discussions recognized a desire not to include `836 in the license and CNS, but if so – just say so with a callout stating that `836 is expressly not part of the license.
I’ve been licensee and licensor in such a situation where a more narrow improvement which was patented was not supposed to be within the scope.
My guess is that the intent of the parties was an “everything in family up/down” type of scope.
People don’t read your posts anymore MM because we have no idea if anything you say is going to be factual or not.
Please try to restore your integrity that perhaps you had 15 years ago.
It’s February and it is the 15 year anniversary of the blight that is Malcolm.
That being said, Malcolm never had any integrity, as his blight traces to the very beginning of his comments.
Does anyone else think this decision is in conflict with the doctrine underlying terminal disclaimers? Example: applicant is rejected with a non-statutory obviousness-type double-patenting rejection in a continuation application. Applicant doesn’t want to be hindered by the linking requirement of a terminal disclaimer, and so “litigates” out the issue (by whatever mechanism–PTAB or courts) and wins. But nope!–continuation is linked anyway by implied license doctrine?
If the patent owner really does not want the parent and continuation linked, this decision provides the mechanism for doing that – expressly state in the license for the parent patent that the continuation is not licensed.
Comments are closed.