Patent Tolls, Not Trolls

Hello all,

Earlier this month, I was invited to attend the annual Bridge Summit of Via Licensing Alliance in San Francisco. Via, which describes itself as the world’s largest patent pool administrator, licenses the patents inside video codecs, wireless charging and, increasingly, AI data centers. The room was full of licensing executives from the US, China, Japan, Korea and Europe, there to settle who collects on the next decade’s technologies. I suspect more than a few of them have been called trolls at some point in their careers. They are not. They run toll booths.

That distinction is the subject of this month’s Feature Story: what happens when licensors stop selling retail and start selling wholesale, who pays to build the road, who owns the booths, and why the AI industry may need more toll booths rather than fewer, particularly when it comes to security.

But first, four shorter takes, and they all happen to be about doors. The Supreme Court bolted one at the PTAB. Unified Patents’ Q3 numbers show where everyone went instead. Senator Grassley wants more light on who bankrolls everyone walking through the courthouse door. And London, after more than a year of deliberation, decided the best SEP reform was mostly none at all.

As usual, while I focus on the macro picture in this newsletter, I want to remind everyone that we track everything happening in this world on a daily basis on LinkedIn, where the breaking news lives.

 Happy reading!

 Louis

__________________

Featured Advisory

Brokerage is what we are known for, but a growing share of our work is advisory, and lately the mix has been broad. We are currently engaged on reverse EoU searches, finding the patents a litigation adversary infringes so our client has something to put on the other side of the scale. We are running target-focused infringement searches and claim charting against named products, freedom to operate analyses, and validity and prior art reviews. We are building damages models for licensing campaigns and conducting IP due diligence for venture investors. And we are helping portfolio owners prune theirs, because not every patent deserves to live forever, and some do not even deserve their next maintenance fee. If any of this sounds useful, write to us at info@tangibleip.biz.

__________________

Settled Expectedly

Unless you’ve been living under a rock, you know the Supreme Court declined on October 19 to hear Google v. VirtaMove, the lead of five petitions challenging the USPTO’s “settled expectations” doctrine. Google brought at least seven amici (I didn’t know Google had so many friends!). The Solicitor General simply asked the Court to pass, and it did, at the first conference after briefing closed. Dennis Crouch expects the companion petitions from Tesla, Intel, FedEx and Kahoot! to meet the same fate. So do I.

Translation: if the patent you want to kill has been on the books for roughly six years or more, forget the PTAB. Institution decisions remain, for all practical purposes, unreviewable under Cuozzo. The door is not just closed. It is bolted from the inside.

Let me be blunt: for patent owners, this is good news…and a relief. A patent that survived years in the market, unchallenged by companies that knew perfectly well it existed, deserves some repose. Infringers who chose to wait and see must now argue invalidity where it belongs, before an Article III court, against a presumption of validity.

You would think I’d be dancing in the street. I tried. My knees and my track record both objected. Readers of this column know my problem was never the number of IPRs. It was the absence of predictability. A doctrine introduced in June 2025, often applied without explanation, and now beyond the reach of any court is not a rule. It is a mood. Today the mood favors owners. Directors change, and so do moods. (Anyone who lived through the PTAB’s “death squad” years remembers how fast the wind turns.)

The Court just left the plumbing to whoever runs the Office. That is why the rule now sitting at the White House matters more than ever. Settled expectations deserve settled rules.

__________________

Back to Marshall

Want to know where the PTAB’s business went? Follow the current. Unified Patents’ Q3 report, released October 16, tells the story in three numbers.

First, district court filings rose 15.6% to 1,166 cases, the busiest quarter since Q2 2025. NPEs filed 650 of them, or 55.7%. The Eastern District of Texas took 400 cases, more than a third of the national docket, and 349 of those were NPE filings. Judge Albright may have hung up his robe, but Marshall never closed for business.

Second, PTAB petitions came in at 106, down 70% from 354 a year earlier. The institution rate did bounce back to 57.3%, but on only 96 decisions, the thinnest sample since the Board’s first full year. A high rate on a handful of survivors tells you who still bothers to file, not that the gate reopened.

Third, the side door we flagged in July and revisited last issue is narrowing. Ex parte reexamination requests fell about 40% from Q2’s record, to 198. Still historically high, but trending down since the Office started screening requests, with 1,354 reexams pending and the median wait up to 68 days from 42 over the prior decade.

What it means: the invalidity fight did not disappear. It moved back to the courthouse, where it is slower, more expensive and decided by people who have to look the patent owner in the eye. For sellers, that changes the math on every portfolio we take to market. For buyers, the window to buy before the assertion cycle restarts is closing fast.

__________________

Follow the Money, Grassley Edition

More courthouse fights need more capital, and on cue the disclosure debate is back. On October 21, the day the federal Advisory Committee on Civil Rules last met, Senate Judiciary Chairman Chuck Grassley sent it a letter urging rules that would “increase public disclosure of third party funding in litigation.” He is also sponsoring the Litigation Funding Transparency Act of 2026. The committee, a decade into the debate, remains undecided.

Let me give the other side its due. Grassley raises a point I take seriously: foreign money using US discovery as a back channel to trade secrets. Nobody in this business should defend a hostile state quietly bankrolling a campaign against a US competitor, and a narrow rule aimed at that would not offend me.

But that is not what is on the table. The proposal backed by Lawyers for Civil Justice and the Chamber’s Institute for Legal Reform would require disclosure of every funder with a financial interest in the case, along with the funding agreement itself. As I wrote in March, disclosing that funding exists is one thing. Handing over amounts and terms gives the defendant a roadmap to how much pain the plaintiff can absorb before it folds. Worse, once disclosure is the rule, silence becomes information. The self-funded inventor might as well wear a sign saying “I can’t afford a long war.” (Imagine a poker game where only one player has to show his cards.)

Taction’s $5.7 billion verdict against Apple, financed as I noted last issue by a Burford affiliate, gave the disclosure lobby the poster child it wanted. Big verdicts make good headlines. Bad rules make for long wars of attrition against the smallest patent owners.

Undecided for a decade, the committee should stay that way.

__________________

London Calling It Off

 

Across the pond, the UK government spent more than a year asking whether it should fix SEP licensing, and concluded, mostly, that it should not.

The July 2025 consultation floated a fast-track court for setting FRAND rates, mandatory disclosure of essential patents, a government essentiality checking service and, the big one, limits on SEP injunctions. Ninety three responses later, the October 13 response dropped nearly all of it, on the theory that the current UK approach is internationally respected. In other words: it isn’t broken, so we won’t fix it. The best SEP reform, it turns out, is the one you don’t make.

What survives is a voluntary, non-binding IPO “FRAND opinion service,” with comments due December 15. An opinion that binds nobody is a thermometer, not a thermostat. Handy if you want to know whether an offer is outrageous. Useless if you actually have a fight.

For patent owners, this is a quiet win. Regulators keep studying SEP licensing, and courts keep doing the actual work, in London, at the UPC, in Munich and now in Brazil. So the side doors are closing on both sides of the Atlantic. Enjoy the leverage while it lasts, and keep the funding agreement in the drawer.

__________________

Feature Story: Patent Tolls, Not Trolls

In 1917, the US Supreme Court had to decide whether a Broadway restaurant owed anything to the composer whose music entertained its diners. Shanley’s argued it was selling steaks, not songs. Justice Holmes had beef with this argument: “If music did not pay, it would be given up.” The composer was Victor Herbert, who three years earlier had co-founded ASCAP with a few fellow songwriters for a simple reason. No composer can chase every restaurant, dance hall and theater humming his tunes. A society of them can. Today ASCAP reports about one million members and roughly $1.8 billion collected in 2024, and nearly every country has its own version.

Nobody calls a songwriter a troll when the bar – an implementer of sorts- pays for the music. Yet the companies that license the technology inside every phone, car and video stream get called exactly that. They are not trolls. They collect tolls. And a century later, the patent world is running the same experiment the music world already finished.

The Toll Road

If you have ever bought a smartphone, you have paid these tolls. You just never got a receipt. Every phone, car and video stream works because competitors agreed on how their products talk to each other, and whoever writes those rules gets paid by everyone who follows them. For 5G alone, more than 60,000 patent families have been declared essential, and the ten largest holders own roughly three quarters of them. The tolls add up to roughly $15 billion a year. Qualcomm’s licensing arm booked about $5.6 billion in fiscal 2025, and Nokia and Ericsson each collect well over a billion more. And the road keeps getting longer: cars, EV chargers, payment terminals, data centers. (Ericsson just sued a payment terminal maker on two continents. Even the card reader at your coffee shop now has a toll booth.)

Who Builds the Road?

Somebody has to pay for all that engineering. Nokia says it has spent around €150 billion on R&D since 2000. There are only two ways to fund that kind of bill. Either the inventor gets paid royalties by everyone who uses the standard, which is the Nokia, Ericsson, InterDigital and Dolby model. Or a company that makes its money elsewhere, on devices, ads or cloud, builds the standard and gives it away, which is how we got Bluetooth, USB and the AV1 video codec.

Take the humble phone charger. Apple did not give up Lightning for USB-C because someone made it a better offer. It gave it up because Brussels told it to, with a deadline. Nothing focuses the mind like a regulation with a date on it. Wireless charging went the other way. Qi is a royalty-bearing standard, the rival camp folded into it, and Apple eventually contributed its MagSafe technology. Competing approaches, one standard, and fewer drawers full of useless cables. When inventors can get paid, they have a reason to show up at the same table.

To be fair, unrealistic royalties can also blow things up. When HEVC Advance launched in 2015 with rates well above the existing pool, the industry built a royalty-free rival instead, and the video world spent a decade with two codec families. Greed is a terrible business plan, even for licensors. But a standard nobody gets paid for still gets written. It just gets written by whoever profits from something else.

Retail, Wholesale and the Jukebox

Regular readers know I distinguish between retail licensing, where a single patent owner tries, often heroically and just as often in vain, to get a large implementer to take a license, and wholesale licensing, where licensing behemoths and patent pools offer one license, one price and patents from dozens of owners at once. Retail licensing is the lone songwriter going bar to bar with an invoice. Wholesale licensing is ASCAP, which sends one bill and never has to sing for it.

In March, in “Swim or Sink! Navigating Patent Pools,” I argued that pools were filling the gaps where one-on-one licensing made no economic sense. Seven months later, the gaps are becoming the market. A Charles River Associates study (paid for, it should be said, by an implementer group) put pools at about 9% of global SEP licensing revenue in 2022. Avanci now covers more than 325 million cars. Nokia, Huawei and Samsung all joined new pools this year, and in September Canon, Sony, Huawei and OPPO pooled more than 125,000 imaging patents. (When rivals who sue each other for a living start singing in the same choir, something has shifted.)

Lest I get carried away by my own metaphor, smartphones are still licensed one deal at a time. The biggest venues negotiate their own terms. Pools own every new road. They do not yet run the whole highway.

Litigate the Whales, Pool the Minnows

Does joining a pool pay? Look at the two ends of the spectrum. InterDigital licenses almost entirely on its own and litigates when talks stall; it reported $834 million in revenue in 2025. Dolby collects mostly through pools, and its licensing business brought in $1.25 billion at a 93% gross margin. (Not bad for a choir member.)

Then look at Verifone. Nokia and Ericsson each sued the payment terminal maker in Europe, and each got a deal. Meanwhile, Nokia helped launch a new Sisvel pool for the long tail of smaller terminal makers that would never justify a lawsuit. That is the real model. Litigate the whales. Pool the minnows.

It is also why pools are not the peaceful alternative to litigation. They depend on it, just as the music license depends on the occasional lawsuit against the bar that thought the jukebox was free. The pool sets the price. The lawsuits are the muscle that make people pay it.

And just as the music societies have lived under antitrust supervision in the US since 1941, the pools are now meeting their regulators. Tesla is taking Avanci’s 5G rate to trial in London, Snap has sued Access Advance and Dolby, and the EU’s new technology transfer rules now apply FRAND to the pool itself. Every collecting society eventually gets audited.

Who Owns the Booths?

Toll booths are assets, and assets trade. The voice codec patents in your 4G and 5G calls were sold in 2018 to an affiliate of Fortress, the same Fortress that just bought IPValue. BlackBerry sold some 32,000 patents in 2023. Dolby bought MPEG LA for $160.6 million, and Via sold its whole HEVC/VVC program to Access Advance for $25 million. Running a pool costs a fortune to build and very little to operate, so whoever ends up owning the booth for a given standard tends to keep most of the toll. Expect more consolidation.

Governments have noticed. Korea pays research teams to turn inventions into standard essential patents. Germany covers up to 70% of a small company’s cost of sitting at the standards table. Even Washington, which quit 66 international organizations in January, kept its seats at the ITU, ISO and IEC. When an administration that allergic to international clubs keeps those particular chairs, the chairs matter.

For sellers, the lesson is simple. A seat in the right pool is now part of a portfolio’s value, and buyers price it that way.

Free Samples on the Next Road

And then there is AI. In April, Anthropic, IBM, Meta, Microsoft and others launched the Shared AI License Foundation, a royalty-free cross-license for AI model patents, with IBM and Microsoft holding more than 90% of the patents involved. Readers of “Thanks a LOT!” will recognize the architecture. Because nothing says “collaboration” quite like a club whose main benefit is that members cannot sue each other.

Free samples have their place. They make the big platforms’ technology the default. What they do not fund is the hard, unglamorous stuff, and AI security is the obvious example. For the labs racing to ship the next model, shared security standards are a speed bump. The specialists who actually invent security fixes have no way to get paid unless they own a platform. A royalty-bearing standard, licensed through a pool, would change that.

There is a precedent. HDCP, the copy protection behind every HDMI connection, is licensed for a fee, and device makers never asked for it. Hollywood simply made it a condition of getting its movies. My prediction: the first paid AI security standard will arrive the same way, imposed by whoever writes the checks. My money is on the insurers.

The Bottom Line

Standards do not write themselves, and they are never free. Someone always pays, either the inventor through royalties or the platform through control. Holmes figured that out in 1917: if music did not pay, it would be given up. Innovation is no different. The bar has paid for the music for a hundred years without calling anyone a troll. The tech industry can learn to pay its tolls too.