The courts have spent a decade telling inventors their ideas are “abstract.” Congress is finally tired of it too.
If you have been following this blog, you already know the story. The Supreme Court’s 2014 Alice decision introduced a vague “abstract idea” exception to patent eligibility that nobody understands. Neither inventors, examiners, Federal Circuit judges, nor apparently not even the Supreme Court itself gets it! This unfortunate ruling has created a patent “black hole.” This black hole has swallowed software patents, diagnostic method patents, AI patents, and large swaths of biotech. Along the way, judges have written increasingly frustrated concurrences asking the Supreme Court to clean up the mess it made.
The Supreme Court has declined every invitation to revisit the issue.
Now Congress has stepped in. On May 1, 2025, the Patent Eligibility Restoration Act (PERA) was reintroduced in both the Senate and House. If it passes, PERA would be the most significant change to §101 since the Patent Act of 1952. For inventors and startups, it is worth understanding what PERA does and why you should not wait for Congress.
What Is Wrong with §101 Right Now?
35 U.S.C. §101 says that patents are available for “any new and useful process, machine, manufacture, or composition of matter.” Four categories; pretty broad. Congress has not substantively changed this language since 1793.
In 2012 (Mayo Collaborative Services v. Prometheus) and 2014 (Alice Corp. v. CLS Bank), the Supreme Court grafted on a two-step “judicial exception” framework. If a claimed invention is directed to an abstract idea, it fails §101 unless there was an “inventive concept”. The problem is that “abstract” is undefined, and the “inventive concept” inquiry overlaps awkwardly with novelty (§102) and obviousness (§103). Courts applied these tests inconsistently, and lower courts followed suit.
The downstream effects have been severe:
- Federal courts routinely invalidate diagnostic patents. These often represented years of genuine research. This has made it significantly harder to patent precision medicine innovations.
- Examiners and Judges routinely invalidate Software and AI patents as “apply it” claims. That is, they argue that the claims take an “abstract idea” and do it on a “generic” computer. These days, pretty much everything, including AI, can be deemed “generic.”
- Hardware and IoT claims that include software-implemented control logic are not immune either. It all depends on how an examiner frames the abstract idea. Some examiners simply highlight the entire claim and assert that it is “abstract!”
- The USPTO and the courts disagree on how to apply the Alice/Mayo framework. The USPTO has the burden of reconciling thousands of inconsistent court rulings into a consistent rule. They formed an internal guidance system that still lets many patents pass. By contrast, the Federal Circuit doesn’t have to worry about consistency. They have moved in a more random but generally more skeptical direction. This divergence creates real strategic uncertainty.
- Every judge on the Federal Circuit has publicly complained that the current doctrine is “incoherent,” “unclear,” or “inconsistent.” Retired Federal Circuit Judge Paul Michel has called the situation an obstacle to domestic investment that only Congress can fix. All 12 active Federal Circuit judges said as much as far back as 2021.
What PERA Actually Does
- Eliminate the judicial exceptions entirely. No more snipe hunts.
- Replace the judicial exceptions with a narrow, statutory list of ineligible subject matter. Under PERA, a claimed invention would only be ineligible if it falls into one of these categories:
- A mathematical formula not part of a claimed process, machine, manufacture, or composition of matter
- A mental process performed solely in the human mind
- An unmodified human gene as it exists in the human body
- An unmodified natural material as it exists in nature
- A process that is “substantially economic, financial, business, social, cultural, or artistic.” However, if the process requires the use of a particular machine, it is OK.
- The last category concerns pure business-method patents. The genomics exclusions address the Myriad Genetics line of cases.
- Decouple §101 from §§102, 103, and 112. This is arguably PERA’s most technically significant provision. Currently, the Alice/Mayo “inventive concept” inquiry confuses eligibility with both novelty and obviousness. This confuses everyone. PERA would end this confusion. Judges and examiners will determine eligibility “without regard to” novelty and obviousness. Questions of novelty and non-obviousness will be evaluated only under §§ 102 and 103.
Who Benefits Most?
Biotech and Diagnostics
This sector has arguably suffered most under the current regime. Mayo effectively created a presumption that correlations between biological phenomena and clinical outcomes are patent-ineligible “laws of nature.” This happens even when it takes major inventive effort to discover the correlation. This has created a perverse incentive: you can patent the kit used to measure a biomarker. But you often cannot patent the diagnostic conclusion that your years of research established.
Under PERA, diagnostic methods that require a specific measurement technique, providing clinically actionable results for specific patient populations, are OK. Examiners and Judges would assess novelty and obviousness questions on their merits under §§102 and 103.
Note: PERA has significant opposition from the medical community on exactly this point. Over 90 medical associations and patient advocacy groups have argued that restoring patent eligibility to biomarkers and diagnostic associations would monopolize life-saving tests and raise healthcare costs. This is a real policy debate, not a fringe concern. Whether PERA’s §102/103 safeguards are sufficient to prevent overbroad diagnostic patents is one of the key disputes holding up the legislation.
Software and AI
PERA would eliminate the “abstract idea” category entirely. This does not mean every software patent would automatically issue — you still need novelty, non-obviousness, adequate written description, and enablement. But it removes the medieval logic rules that currently allow examiners and litigants to kill software claims in the cradle.
For AI inventors specifically, PERA would resolve the tension between the differing USPTO and Federal Circuit standards. The question would shift from “is this abstract?” to “is this novel and non-obvious?”
Hardware and IoT
The direct benefit here is more modest, since hardware inventors face Alice challenges less frequently. The real benefit is indirect: investor-side clarity. Startups raising seed or Series A rounds routinely face due diligence questions about patent eligibility. PERA would significantly reduce the legal uncertainty that clouds those conversations. These will help any hardware product with embedded software control logic.
The US vs. China and Europe
This is the geopolitical hook that has given PERA its bipartisan traction. Europe and China routinely grant Software, AI, and diagnostic patents that the US rejects. American inventors are at a disadvantage because their foreign competitors have stronger patent protection. PERA’s sponsors have repeatedly invoked this asymmetry, and it resonates in the current policy climate.
PERA Is Not Law Yet
Don’t hold your breath that Congress will pass PERA this year. The bill has been introduced several times. It has bipartisan support in both chambers. Multiple former USPTO Directors — including Andrei Iancu (2018–2021) and David Kappos (2009–2013) — have publicly backed it. So have retired Federal Circuit judges.
And yet PERA keeps failing to pass. Why?
Some industries simply don’t like patents at all! For example, the medical community’s opposition is organized and vocal. They worry that PERA might allow patents on unmodified DNA. PERA’s drafters have tried to address this with the statutory exclusions for unmodified human genes and natural materials, but critics argue those carve-outs are narrower than the judicial protections they replace.
There is also the pharmaceutical industry’s “pull up that ladder” position: large pharma tends to favor strong patent rights. But they are wary of patents held by biotech startups and academic medical centers. Biotech startups, conversely, tend to strongly favor PERA.
The bottom line: PERA may pass in this Congress. The political environment — US competitiveness with China, a pro-patent USPTO directorship under Director Squires — is probably more favorable now than it has been in years. But “probably more favorable” is not “certain.” Patent strategy cannot be built on pending legislation.
What to do at present (before Congress passes PERA)
- Draft claims that survive the current law, but that might be strengthened under PERA.
This is not as contradictory as it sounds. Try to articulate what is technically improved in your system. This can include the specific architectural choices, the particular algorithm design, the novel application of a known technique to a problem it has never been applied to before. This kind of specificity satisfies the Federal Circuit’s post-Recentive demand for disclosed improvements and positions you well under PERA, where novelty and non-obviousness are the operative questions anyway.
- Do not treat §101 issues as fatal before you get an examination.
Examiners and Judges apply the Alice/Mayo doctrine inconsistently. USPTO examiners and the courts do not always agree, and some art units at the USPTO are significantly more hostile to software and AI claims than others. Prosecution strategy matters enormously. This includes giving examples and alternative embodiments in the specification. Consider filing continuations and continuations-in-part to keep patent applications alive for longer. Office action response strategy can also determine eligibility outcomes for nominally similar inventions.
- Consider a provisional application
If you are in the early stages of an AI, diagnostic, or software-based invention, filing a provisional application establishes your priority date immediately. You get twelve months to watch how PERA and the case law develop before committing to final claim language. Given that the USPTO’s examination backlog was well over 800,000 applications in early 2025, that provisional year is practically free time.
- For diagnostic and biotech inventors specifically: file both composition and method claims.
Even under current law, composition of matter claims for novel biological tools, reagents, and test kits are generally more survivable under §101 than method claims for their application. Filing a robust application that covers both the tool and the method gives you options — and potentially valuable claims — regardless of what Congress does or does not do with PERA.
Conclusion
Section 101’s current state is, in the polite words of the Federal Circuit, “incoherent.” PERA is the most serious congressional attempt to fix it in the 12 years since Alice. Whether it passes in 2026 or eventually dies again in committee, it tells you something important about the direction the law is heading.
If your invention involves software, AI, diagnostics, or biotechnology, now is exactly the wrong time to wait and see. The prior art clock does not pause for pending legislation, and your competitors are filing applications today. The right move is to build a patent strategy that works under the current law — and that we can adjust if and when the law changes.