Buckle up for another dry post, everyone! I figured an update was in store on the Gogo litigation due to recent important events. Notably, there are no imminent litigation events which are likely to cause trading opportunities one way or another. The trial date for this litigation isn’t until 2025, and Smartsky, as the newcomer (or perhaps more appropriately, the laggard) in the business aviation space, has almost no incentive to settle any time soon. Nor does Gogo have incentive to offer anything remotely enticing to Smartsky to encourage settlement.
Recall Smartsky is suing Gogo for infringement of 4 patents: United States Patent Nos. 9,312,947 (“the ’947 Patent”), 11,223,417 (“the ’417 Patent”), 10,257,717 (“the ’717 Patent”), and 9,730,077 (“the ’077 Patent”).
As a brief background on the patents, the ’947 and ’417 patents relate to a communications “mesh network” which operates to provide continuous communications service to aircraft. The ’717 and ’077 relate to the way in which the RF beams are formed and projected to aircraft. Notably, the ’947 patent and ’417 patent share a specification, meaning these are related patents. Further, though not sharing identical specifications, the ’077 and ’717 patents share some of same figures.
You may also recall that in my first post on Gogo, we discussed why Smartsky’s motion for preliminary injunction was likely to lose. In order to prevail, Smartsky had to show that (1) it was likely to win the case on the merits and (2) that failing to institute an injunction would cause Smartsky irreparable harm. There are actually 4 factors in deciding whether or not to issue an injunction (the other two being balance of hardships to the parties and the public interest) though it is generally accepted that the first two factors must be present. In other words, Gogo would win if it created doubts as to validity of the patents or infringement of the patents, created doubts as to whether or not Smartsky would actually be harmed in a way that future monetary damages would not compensate, or some combination of both.
Smartsky did end up losing the motion for PI, but it’s probably not worth a victory lap—preliminary injunctions are pretty hard to get. Lex Machina looked at all motions for preliminary injunctions in patent cases from 2013 through 2020 and found that only 41% of preliminary injunctions were granted. The number is higher when considering practicing entities’ motions versus non-practicing entities’ motions (aka “patent trolls”); this makes obvious sense because it’s hard to argue an alleged infringer will cause you “irreparable harm” when you don’t actually make anything. Still, while Smartsky is a practicing entity (they will offer a product for sale which employs the technology in the asserted patents), the case seemed pretty weak, particularly given Smartsky’s struggles to get its network up and running.
Unhappy with the District Court’s decision denying the preliminary injunction, Smartsky immediately appealed the decision. Typically, mid-litigation appeals (“interlocutory appeals”) are not procedurally guaranteed; however orders on injunctions are a special type of order and are appealable as a matter of right. So Smartsky, not wanting to miss any opportunity to throw a wrench in Gogo’s 5G release, took advantage of this right. The parties briefed the appeal, and the hearing at the Federal Circuit took place this past Monday in front of Judges Chen, Hughes, and Cunningham (recall from my last post on LQDA that Fed. Cir. judges hear these cases in panels of 3).
I’m not going to get into every reason the District Court found Smartsky’s arguments insufficient, nor am I going to go through every element of the appeal. After all, I expect the Federal Circuit to simply affirm the District Court, and predicting simple things is neither interesting nor impressive. Just listen to this brief exchange during the Fed. Cir. hearing last week where Judge Cunningham makes note of Gogo’s pathways to success, described above:
What I do want to talk about is the way Smartsky has shown its hand with respect to the rest of the District Court case during the course of this appeal.
PATENT LAW EDUCATION: CLAIM CONSTRUCTION, PRELIMINARY INJUNCTIONS, AND REPLY BRIEFS
In order to fully get into this, we have to do a little more patent law education. I’ve talked briefly in the past about claim construction, and my last post on LQDA shows just how important claim construction can be to the outcome of a case. It’s a major reason why so many patent cases settle after a Markman hearing. Of the roughly 11.5% of patent cases that even make it to claim construction stage, over half end up settling some time between the Markman and trial.
So what exactly are we doing talking about claim construction in this Smartsky v. Gogo case anyway? We haven’t had a Markman hearing.
It turns out we have had a claim construction ruling—just not a formal, binding one. As we discussed above, Smartsky had to show that it was likely to win on the merits in order to obtain a preliminary injunction. Given the critical nature of claim construction to the outcome of the case, it would be ridiculous to think we could divine the outcome with any accuracy absent at least some sort of agreement on what the claims mean.
Prior to the preliminary injunction ruling, the parties offered up technical experts and engineers who were deposed and who gave sworn testimony saying “this is what these claims mean” and “this is what Gogo’s network does” (or in the case of Smartsky’s expert, “this is what we think Gogo’s network does”). It’s all a little technical and, again, not necessary to delve into every detail.
What is important is that, after Smartsky took the deposition of Gogo’s technical expert Dr. William Michalson and Gogo’s engineer Mr. Yong Liu, Smartsky realized it had a problem: the way it had characterized its patent claims in its complaint and earlier filings was pretty clearly inapposite with the way Gogo’s network actually functions. Smartsky would have to change the interpretation of its claims in order to legitimately claim that Gogo’s network infringed. Smartsky would tell you—and the District Court agreed—it was simply supplementing the claim interpretation, but for some reasons we’ll get into in just a bit, this supplemental rationale actually has some huge implications.
You might be thinking “That’s deeply unfair! You can’t just change your theories of infringement to suit your needs as the case develops!”
You’d be right. Generally speaking, you can’t add new information into reply briefing and then claim victory when the other party hasn’t had a chance to meaningfully rebut your claims. Contrary to literally every tv show and movie ever, trial by surprise is not an acceptable method of litigation.
But like it or not, the District Court allowed Smartsky to amend its reasoning and alter the claim interpretation. We could spend all day discussing whether or not that was the right call, and whether or not Smartsky’s reply brief was actually responsive to Gogo’s Answer, but the simplest explanation for it is that the earlier you are in the litigation, the more lenient a judge will be when it comes to amending your reasoning. After all, there’s still plenty of litigation time left before we impanel a jury, so it makes some sense to give Smartsky the benefit of the doubt.
It’s also very important to note that Smartsky is not bound by the claim construction it has set forth at this stage. We could easily see Smartsky propose different claim interpretations once we do actually get to the Markman stage.
So why does it matter at all then?
It matters because we know why Smartsky had to amend its reasoning, and it matters because the best argument Smartsky has come up with with respect to patent infringement at this stage creates substantial issues of patent validity. I.e., the best arguments for why Gogo infringes actually hand Gogo the tools it needs to invalidate the patents at issue.
SMARTSKY’S CONUNDRUM: ISSUES WITH THE ’947/’417
Let’s first take a look at the ’947 and a’417 patents together which, given the identical nature of the respective specifications, suffer from the same issues.
Both patents describe a network which has a radio providing a communication link that is “continuous and uninterrupted in time.”
’947 Claim 1:
’417 Claim 1:
This purpose of this claim seems fairly self-explanatory at first glance—a steady connection that doesn’t get dropped is important for users and seems like a good goal for a communication network. But thinking back to my last Liquidia piece, we again quickly realize we can be pretty persnickety about this. Does “continuous and uninterrupted in time” mean a steady signal? Does it mean the user doesn’t experience any breaks in service, but that small, unobservable breaks are acceptable? The Federal Circuit, noting that the patent describes signal connections and does not, in fact, reference an application layer of any sort, indicated that the “continuous” nature of the link must be some technical, physical link of the signal.
However, during the course of initial discovery (usually discovery doesn’t happen so quickly, but when you move for a preliminary injunction some discovery is needed at the outset to rule on things like “likelihood of winning on the merits,” etc.) Smartsky learned that Gogo doesn’t actually accomplish its network meshing by having its on-board radios connect to multiple towers at the same time (a “soft handoff”). Rather, the Gogo enabled-plane will physically disconnect from the first base station signal before connecting to the net base station signal (a “hard handoff”). The user is unlikely to experience any noticeable hiccup in service, but at least momentarily, the connection is physically lost.
So Smartsky amended its pleadings to expand the interpretation of the data connection, saying that the claim reads on the “data communications link generally,” which would include both the physical layer (i.e., the signal) and the application layer (i.e., the user experience), thus capturing both “soft” and “hard” handoffs. However, this interpretation appears to be inconsistent with the inventor’s intent. Here’s Gogo’s attorney describing the prosecution history of the ’947 and ’417 patents:
Smartsky counters Gogo’s arguments by saying that claim scope limitations must be present in a clear disavowal made by the inventor. In other words, we don’t exclude embodiments from our claim interpretation simply because the inventor left one out. We take the claims for gospel, and we assume anything which fits the definition of the claim is included in the scope of the claim unless the inventor tells us something shouldn’t be there. After all, it’s usually impossible to include every embodiment, and the mere fact that an inventor hasn’t listed every embodiment shouldn’t mean that the claim scope shrinks. By way of example, if I were to draft a patent claim for “a food comprised of a carbohydrate shell and meat interior, wherein the carbohydrate shell surrounds at least 3 sides of the meat interior,” we should theoretically assume I meant both hot dogs and tacos, unless I tell you in the patent or in the back and forth with the patent office that I didn’t actually mean for my patent to cover hot dogs.
However, we also know from a 2016 Federal Circuit case (Poly-America, L.P. v. API Industries, Inc.) that, while disavowal must be clear, it need not be explicit. We also know from M.I.T. v. Shire Pharmaceuticals, Inc. that courts will look to prosecution history, which would include the claim chart Gogo’s counsel describes in the audio excerpt, above. And though the claim limitation of hard vs. soft handoffs was not under scrutiny by the USPTO at the time Mr. Alcorn made his apparently distinguishing statements (as was the situation in the MIT case), it seems highly likely that these these circumstances, taken in whole, will lead to a favorable claim construction for Gogo on the ’947 and ’417 patents.
But even if Smartsky gets favorable claim construction later on, this isn’t the only issue with these patents. Recall again from the last Liquidia post that patents must be sufficiently described and enabled under 35 U.S.C. §112 to be valid. If the patent doesn’t actually describe the invention, or if a POSA can’t read the patent and understand how to practice it without “undue experimentation,” the patent is invalid.
So what does §112 have to do with the current patents?
Data communication of the sort described in the patent—the type of data communication like wifi and bluetooth where you want things to be consistent so that devices can all talk to each other, regardless of manufacturer—is governed by wireless broadband standards created by the Institute of Electronics and Electrical Engineers (IEEE). One of these standards is “Standard 802.16” otherwise known as WiMAX (standing for "Worldwide Interoperability for Microwave Access"). Another, perhaps more recognizable type of these standards is the “LTE” standard.
When the inventor, Donald Alcorn, was describing this patent, he made sure to include reference to these standards which were compatible with his invention. That makes sense, as inventors want to disclose several “embodiments” which are encapsulated by their invention so as to show that they know what they’re talking about.
But Mr. Alcorn (or far more likely, his patent agent or patent attorney) seem to have made a bit of a mistake. Somewhere within the many thousands of pages of text that describe the 802.16 and LTE standards, there is a description of “hard handoffs.” They do not, however, describe “soft handoffs.” Perhaps realizing this, Mr. Alcorn’s patent agent or patent attorney included a few catch-all phrases. The standards might need to be modified in order to practice this invention.
Indeed, “all described embodiments may utilize . . . modified versions of the next generation wireless technology.”
What the inventor likely wanted to say was “This invention applies to 802.16 and LTE standards. It might also apply to future versions. I want to future-proof this patent, so minor updates should be included when interpreting this thing!”
But that isn’t what he said. He instead said something more along the lines of, “All of this stuff might only work on modified versions of the standard.” He then went on to say . . . nothing. There was no additional description of what modifications might be required to actually practice the invention, leaving us (and likely a POSA) unable to discern whether or not this invention works without—you guessed it—undue experimentation.
SMARTSKY’S CONUNDRUM: ISSUES WITH THE ’717
Moving on to the ’717, we have perhaps even more egregious, and fatal, issues with Smartsky’s proposed interpretations.
The ’717 patent describes a communication network with two base stations: one using licensed spectrum and one using unlicensed spectrum. A radio on the aircraft is able to jump from the first base station to the next to create a “handover” effect between the two types of base stations.
’717 Claim 1:
In the spirit of continued persnickety-ness, we can acknowledge there may be some ambiguity in how the base stations function. Is it that the first base station only employs unlicensed spectrum?
It’s fairly clear from the intrinsic evidence in the specification that the inventiveness of this patent arises from the fact that we have a transition from unlicensed to licensed spectrum. Here’s a brief excerpt from the background of the invention describing traditional “rigid” networks, before going on to explain the more flexible cross-network mesh:
It’s also pretty clear that Smartsky originally interpreted the claim to require a single radio which performs the connectivity function. Here’s an excerpt from their complaint wherein Smartsky describes why Gogo’s network infringes:
But again, during the course of initial discovery, Smartsky learned something else about the Gogo network: Gogo does not use a single radio on its planes which can communicate with licensed and unlicensed towers. Instead, it uses two separate radios. Rather than switching a single on-board radio from one type of tower to the next, each radio operates in parallel, searching for towers of the type that it can communicate with. If the radio finds such a tower, it checks via on-board computers to see if it is receiving the strongest signal, or if the other radio is receiving the strongest signal, and then decides which channel to use based on the relative information.
So Smartsky amended its pleadings with this information. The patent claims—according to Smartsky—do not require a single radio. Nor, according to Smartsky, must the base stations only employ the type of spectrum listed in the claim. Smartsky interprets the patent claim to include a plurality of radios, a first base station which employs at least unlicensed spectrum, and a second base station which employs at least licensed spectrum.
Smartsky’s new interpretation creates the best bet for Gogo infringement. And regardless of whether or not this interpretation is the right one, Smartsky’s newly stated position creates some issues for it later on. If Smartsky’s interpretation is accepted as true, the only “inventive” concept in the patent at all is that the first base station—the one employing unlicensed spectrum—is “pointed at the horizon.” Smartsky says this is novel because it allows a plane at greater distance to receive unlicensed spectrum without ground interference from devices running on that same spectrum. But as Gogo’s counsel noted at the Federal Circuit hearing, this wasn’t actually inventive at all:
By seeking an interpretation of the ’717 claim that creates infringement, Smartsky has inadvertently created an interpretation that renders the patent completely anticipated by the prior art. Here’s just one example of many other patents pre-dating the ’717 patent showing base stations directing the radiation pattern toward the horizon:
The Federal Circuit panel was similarly unimpressed, calling the majority of the claim “pretty basic,” indicating that Smartsky may have to change its position on the ’717 patent claim interpretation if it wants the claim to survive this litigation:
SMARTSKY’S CONUNDRUM: ISSUES WITH THE ’077
Call it laziness or efficiency, but I’m leaving off discussion of the ’077 for now. The ’077 was not briefed in the appeal of the injunction, and likely for good reason: it seems to be the weakest of the 4 patents (which is surely saying something). Judge Williams (the D. Del. judge presiding over the case) called the parties’ briefing of the ’077 “scant” before denying the preliminary injunction. It’s probably too early to call, but I wouldn’t be surprised if this patent got dropped from the lawsuit before trial.
INCREDULITY ON IRREPARABLE HARM
It’s worth briefly addressing the Federal Circuit’s comments on the asserted irreparable harm that Gogo’s alleged infringement creates. Smartsky claims that, while lost sales of equipment revenue can be calculated at trial, the decision to go with Gogo 5G over Smartsky creates unknowable service revenue losses. It’s not a particularly compelling argument. Lost sales are lost sales, regardless of whether or not those sales are one-time or subscription sales, and the uncertainty around adoption date—the idea that customers are holding out for an indeterminate amount of time for Gogo 5G and creating unknowable harm to Smartsky—is not a particularly convincing one. Smartsky counsel makes the argument that contract lengths are 17 years, which only serves to reinforce the insignificance of the adoption dates. With such long contracts, the few months of “holdouts” (if any truly exist) who ultimately go with Smartsky but who could have chosen Smartsky sooner is immaterial.
The Federal Circuit did not sound convinced:
Remember, for injunction purposes, this alone should be sufficient for Gogo to defeat Smartsky’s appeal.
FINAL THOUGHTS: WHAT’S IN STORE FOR THE MARKMAN?
So will Smartsky stick to its guns now, or will we see another change in reasoning ahead of the Markman hearing?
I would bet it Smartsky is done with major claim construction modifications, though it will probably come down to jury strategy. It’s usually easier for a jury to grasp concepts around infringement than it is for a jury to grasp concepts of validity. If I’m a defendant, I want to be able to tell a jury “I don’t infringe because my thing isn’t the same as this patent claim.” It’s a tougher row to hoe to tell a jury “this patent isn’t valid because the patent doesn’t disclose sufficient detail to be enabling” or “these two things that were done before this patent was filed and which qualify as prior art actually would have been enough for a POSA to understand they should have been combined to create this patented technology.” On the other side of the coin, if I’m a plaintiff, I want to be able to say “the other side has come up with this really complicated argument around validity, but the USPTO gave us the patent—surely that’s worth something! Oh and by the way, our very broad patent claims clearly cover the defendant’s technology.”
As a reminder, I am long . I have not made any changes to the position. Given the recent press releases around global broadband and the continued development of the ESA antennae which can be fitted to pretty much all aircraft, I expect Gogo to be in a great position for years to come as they finish rolling out 5G and expand beyond the North American offering. I also expect a strong upward revision of the FCF numbers for 2025 and beyond, but don’t expect new numbers in the immediate future.
I quite literally stumbled onto your work as I continued my own work on the spirit/jetblue deal, and since then I've been on a bit of a valorem binge.
Gogo had never been on my radar, but I've gained such respect for the depth and quality of your writeups I moved it to the top of my pile this week. Obviously very early into familiarizing with the businesses history and current situation but quickly becoming, at minimum, intrigued. I've spent the past couple of days starting the dive into their filings, call transcripts, and one particularly good business breakdowns episode.
Right out of the gate I was struck by their licensed spectrum asset (which they paid a remarkably low $30ish million for if recall serves me well).
I have virtually no knowledge of this assets value today, but with thought given to spectrum auction "wins" by Verizon and the like in recent hears, one could easily hypothesize a much larger value. Granted, different bands, different uses, different market application sizes etc make for only very loose comparisons.
On the other hand, given they are basically unique in posessing this air-to-ground asset, I could be convinced it is truly a hidden gem if business air connectivity really goes to total penetration with good quality expectations over the next decade.
Has your other valuation work on gogo led you to any thoughts on what that spectrum would be worth to market bidders in today's world?
Could be that the operating metrics of the business far overshadow any attempt to pin a price range on the ATG spectrum, but it's a brain itch that needs scratches to some extent for me.
Apologies for long and somewhat rambling post. A bit eye weary from first couple days back on the grind after a different sort of binging last week (Christmas sweets, etc of course).
Appreciate the work you put into these write-ups. At bare minimum they are eye opening to me that I should be more careful in attempting to assess certain legal situations myself. Expert "consultation" is undervalued here in my opinion.