At the Senate Judiciary Committee's first full-committee hearing on the Patent Eligibility Restoration Act, opponents defended the Alice/Mayo framework as a shield against “patent trolls.” That defense fails on the law, fails on the evidence, and misidentifies the doctrine's actual victims.

On July 14, 2026, the Senate Judiciary Committee held its first full-committee hearing on S. 1546, the Patent Eligibility Restoration Act of 2025. The Tillis-Coons bill would replace the judge-made exceptions to 35 U.S.C. § 101 with a defined list of statutory exclusions and leave Sections 102, 103, and 112 alone. A&O Shearman's IP Litigation blog published a careful summary of the hearing, reporting testimony from former USPTO Director Andrei Iancu, the Alliance for Aging Research's Sue Peschin, the Computer & Communications Industry Association's J. John Lee, and Dr. Debra Leonard, and noting Chairman Grassley's remark that the Committee is still “studying and trying to understand” the bill.

As reporting, the note is fair. But it relays a claim that got far less resistance at the hearing than it deserved: CCIA's assertion that Section 101 guards against abusive litigation by patent assertion entities, and that fixing eligibility law would let the trolls loose. That framing has run the eligibility debate for a decade. It is wrong, and patent owners should say so.

Eligibility is a subject-matter rule, not an abuse remedy

Start with what Section 101 actually does. It asks one threshold question: is this invention the kind of thing the patent system covers at all? It does not ask who owns the patent. It does not ask whether the owner manufactures anything, how the patent is asserted, or whether the assertion has any merit. A doctrine that cannot see the difference between a shell company and a university cannot work as an abuse filter. Stretched into that role, it destroys valid rights and invalid ones without telling them apart.

The Patent Act is not short of tools aimed at weak patents and bad behavior. Sections 102 and 103 eliminate claims that are old or obvious. Section 112 polices vague and overbroad claiming. Inter partes review exists so implementers can test validity at the Patent Office for a fraction of litigation cost. Rule 11 reaches frivolous filings, and after the Supreme Court's decision in Octane Fitness, fee-shifting under Section 285 became a real deterrent against abusive plaintiffs. Each of those mechanisms targets the conduct or the claim defect that makes an assertion abusive. Section 101 targets neither. Deploying it as troll control means choosing the one instrument in the toolbox that cannot distinguish the patents we want from the patents we don't.

PERA's critics rarely engage with this point. Engaging with it concedes the game. If the concern is litigation abuse, the answer is abuse remedies, and the only work Section 101 does that Sections 102, 103, and 112 cannot do is kill patents that are novel, nonobvious, and fully described. That is not a loophole the trolls exploit. It is the feature the doctrine's defenders are fighting to keep.

Look at who Section 101 actually kills

If Alice and Mayo were troll-hunting instruments, the casualty list would be full of shell companies. It is full of operating companies and research institutions instead. The hearing record makes the point on its own.

American Axle involved a driveshaft manufacturer whose method for reducing vibration in automotive shaft assemblies was held ineligible as directed to a law of nature. In the Federal Circuit's 2020 en banc proceedings, members of the court warned that its eligibility rulings had grown so inconsistent and unpredictable that they threatened the innovation incentive in every field, and that the victims ran well beyond one inventor to the national interest in an industrial economy. The Solicitor General recommended that the Supreme Court grant review and fix the standard. The Court denied review in 2022 anyway. A parts maker asserting a manufacturing method against a competitor is nobody's definition of a troll.

In Ariosa Diagnostics v. Sequenom, the Federal Circuit invalidated patents on a non-invasive prenatal test. Judge Linn wrote separately, as Ms. Peschin recounted at the hearing, to say the invention was exactly what the patent system should encourage—and that Mayo's sweeping language compelled him to strike it down regardless. He called on Congress to fix the law. Mr. Iancu's testimony added a diagnostic for myasthenia gravis condemned as a “law of nature” despite enabling a previously impossible diagnosis, and a Duchenne muscular dystrophy gene-therapy patent invalidated in 2024 before the Federal Circuit reversed in 2026. He testified that diagnostics investment fell by more than $9 billion after Mayo, and that eligibility unpredictability lands hardest on startups and small inventors, while large incumbents fall back on scale, secrecy, and litigation budgets.

None of those patent owners was a patent assertion entity. Each held technology that was concededly new. Each lost, not because the invention was old or obvious or inadequately described, but because a judge-made exception swallowed it. A doctrine whose leading casualties are a driveshaft maker, a prenatal-testing innovator, and a gene-therapy developer is not protecting anyone from trolls. It is transferring value from inventors to implementers and calling the transfer hygiene.

The $30 billion number comes from the buyers' side of the market

The empirical anchor of the anti-troll defense, repeated in the CCIA testimony as an estimated $30 billion in annual costs before Alice, traces to a 2012 study by James Bessen and Michael Meurer, later published in the Cornell Law Review, estimating $29 billion in “direct costs” from NPE assertions in 2011. Anyone invoking that figure in 2026 should have to defend where it came from and how it was built. Both are compromised.

Take provenance first. Bessen and Meurer built their estimate on a survey distributed by, and a litigation database supplied by, RPX Corporation—a publicly traded defensive patent aggregator whose subscription business depends on corporate fear of patent assertions. The headline statistic in the case for treating Section 101 as troll control was manufactured with data curated by a company that sells troll defense. Courts exclude interested-party evidence far less compromised than that.

The method fares no better. Professors David Schwartz and Jay Kesan showed that the study counts settlements, licenses, and judgments as deadweight “costs,” when economists classify those payments as transfers—the intended rewards of the patent system—and that it measured the benefits side of the ledger from the SEC filings of just ten publicly traded NPEs, while defining the NPE category broadly enough to sweep in universities and individual inventors. RPX itself later undercut the number. Its own 2014 cost report put NPE-related legal fees, settlements, and judgments at $12.2 billion, less than half the figure that still circulates in congressional testimony.

A statistic sourced from one side of a two-sided market, computed by counting the other side's lawful compensation as social waste, and disowned in scale by its own data supplier is not a foundation for national innovation policy. It is advocacy wearing the costume of economics.

“Troll” is an epithet, not an economic category

The deeper defect in the anti-troll framing is definitional. As then-FTC Commissioner Joshua Wright observed, the non-practicing-entity category includes universities, startups, semiconductor design houses, individual inventors, and established companies that develop technology without manufacturing the end product. Licensing is not a pathology of the patent system. It is one of the system's intended functions—the mechanism the USPTO credits with moving federally funded research out of laboratories and into commerce. A doctrine calibrated to punish assertion by non-manufacturers punishes the entire supply side of the technology market.

The procedural reality makes it worse. Alice invalidations are routinely obtained on the pleadings, before claim construction, before discovery, before any evidence is heard, at the cheapest possible moment for the best-funded party in the case. That asymmetry does not fall on trolls, who price motion practice into their business model. It falls on the small patent owner whose entire enforcement case must survive an abstract-idea argument advanced by a defendant with a nine-figure litigation budget and decided on a cold record.

The market has already priced this in, and the government has documented it. GAO's December 2024 report on patent litigation funding, GAO-25-107214, records what universities and inventors told the agency: fewer law firms will take patent cases on contingency because of the costs and risks unique to patent litigation, and resource-constrained owners cannot file at all without third-party capital—capital that is itself rationed by the risk that a duly issued patent will be invalidated after full investment in the case. Eligibility uncertainty operates as a discount applied to every patent owner's asset before a single infringer is ever confronted. The beneficiaries of that discount are not the public. They are the companies practicing the technology without paying for it.

The opposition's own statistics prove the patent owners' point

The CCIA testimony leaned on an Iowa Law Review study finding that from 2012 to 2023 the Federal Circuit affirmed 85.3 percent of district-court Section 101 decisions and 95.5 percent of USPTO Section 101 decisions, with dissents in only 6.5 percent of appeals. This was offered as proof that the doctrine is stable and administrable. Set it beside the fact, cited in Mr. Iancu's testimony, that by 2021 every one of the twelve then-sitting Federal Circuit judges had publicly lamented the incoherence of eligibility law. Both things cannot signify clarity. A court that affirms nearly everything under a test its own members call incoherent is not applying settled law. It is deferring to outcomes under a standard nobody can articulate, and high affirmance of a one-way ratchet is what a stacked deck looks like from the appellate docket.

The prosperity evidence suffers the same flaw. Nearly 50,000 U.S. AI patents granted in 2023 and three-quarters of global AI venture capital flowing to American firms tell us about the sectors Alice spared. They tell us nothing about the sectors Alice gutted. The $9 billion decline in diagnostics investment after Mayo, the withdrawn Alzheimer's imaging partner and the abandoned melanoma research Ms. Peschin described—none of that shows up in AI venture statistics, and no one should pretend it is answered by them. The question was never whether American innovation survives Section 101. The question is who was sacrificed to it. The record says the sacrifice was made in medical diagnostics, in biotech, and in the workshops of small inventors, while the patent assertion entities the doctrine was supposedly aimed at adapted and kept filing.

PERA keeps every safeguard that actually targets abuse

The final answer to the anti-troll defense is the bill's own architecture. PERA leaves Sections 102, 103, and 112 exactly as they are. Every claim that is old, obvious, vague, or undescribed remains invalid. Inter partes review remains available. Fee-shifting and Rule 11 remain available. What disappears is the shortcut: the ability to erase a concededly novel, nonobvious, fully disclosed invention on the pleadings by calling it an abstract idea or a law of nature. If the opposition's concern were genuinely litigation abuse, losing that shortcut would cost it nothing, because abuse is defined by claim defects and litigation conduct that other doctrines fully reach.

Dr. Leonard's testimony deserves a different response, because it raises a drafting question rather than a framing one. Her concern is that excluding only an “unmodified human gene, as that gene exists in the human body” could be read to re-patent genes through routine purification. That is the kind of objection the legislative process exists to resolve, and the sponsors treated it that way, inviting her to propose language consistent with the shared goal of codifying Myriad. A fixable definition is an argument for markup. It is not an argument for preserving Alice and Mayo.

Chairman Grassley says the Committee is still studying the bill. Fair enough. The study should begin by retiring the premise that has distorted this debate since 2014. Section 101 was never an anti-troll statute. It has been used as one, and the record of that use is a decade of destroyed diagnostics, dead gene therapies, discounted portfolios, and small inventors priced out of their own enforcement rights, while the entities the doctrine was supposedly aimed at kept filing. Congress does not need to choose between fixing eligibility law and policing litigation abuse. The Patent Act already does both, everywhere except in the judge-made exceptions PERA would finally remove.

Ramey LLP is a Texas-based intellectual property law firm dedicated to representing small patent owners, startups, and independent inventors in disputes against larger corporations.