Beyond Section 101: 2026 Intellectual Property Outlook
A blog cannot live on patent eligibility alone. Below are the intellectual property developments we're tracking closely and why they matter.
Section 101 in 2026: A Deliberate Split Between Examination and Enforcement
You are reading the Section 101 blog, so this must go first.
The U.S. Court of Appeals for the Federal Circuit's Recentive decision reinforced a familiar message for artificial intelligence (AI) and software claims in litigation: Applying generic techniques to a new data environment, with results‑oriented claim language, is ineligible under Section 101 absent technological improvement to the model or computer itself. At the same time, the U.S. Patent and Trademark Office's (USPTO) Desjardins framework and subsequent guidance nudge examiners to treat Section 101 as a gatekeeper and address breadth, inventiveness and clarity under Sections 102, 103 and 112, recognizing practical applications at Step 2A where appropriate.
We are watching whether this divergence hardens in 2026: steadier allowance paths at the USPTO for well‑documented implementations but continued Federal Circuit skepticism for "ends‑not‑means" claims in court.
On the horizon, Congress has again aired PERA and a new wave of commentary, and petitioners continue to seek U.S. Supreme Court intervention – developments that remain important signals even if near‑term change is unlikely.
Patent Trial and Appeal Board, Reexamination and the Rebalanced Forum Mix
Under new USPTO leadership, discretionary denials were reinforced, making Inter Partes Review less predictable and pushing more disputes back into district courts while ex parte reexams tick up.
We're tracking three datapoints: 1) what rationales drive denials, 2) whether petitioners continue pivoting to reexam as the validity tool of choice, and 3) whether – or how quickly – district courts advance merits as stays become harder to secure.
Copyright and AI: Inputs, Outputs and a Supreme Court Inflection
A recently argued Supreme Court case will possibly redraw the contours of internet service provider secondary liability, with immediate implications for notice‑and‑takedown, repeat‑infringer policies and damages exposure across platforms and rightsholders.
At the same time, courts began sorting the training data/fair use debate in 2025, with increasing emphasis on market effect and lawful acquisition; 2026 should bring some early appellate decisions across the various AI training data cases.
We're also watching whether the human‑authorship requirement for registrations is clarified by courts or through U.S. Copyright Office leadership and guidance. The U.S. Department of Justice urged the Supreme Court to reject the pending petition on this issue.
Trademarks and Trade Dress: Use, Function and Who Decides What
Examiners and the Trademark Trial and Appeal Board will likely get more structured "failure‑to‑function" guidance after critiques of conclusory refusals, which should make the evidentiary burden and analysis more predictable at the USPTO.
On the merits side, 2026 appellate activity includes questions about when designs and words on merchandise operate as source identifiers versus ornamentation and how conceptual strength is assessed. The Supreme Court in 2026 already declined to hear a case that challenges the doctrine of "foreign equivalents." All this will all affect clearance, registrability and trial posture.
Trade Secrets: Particularity, Limitations and AI‑Centric "Secrets"
Courts continue to demand early, particularized identifications and real, operational safeguards before allowing broad discovery. That trend is intensifying. Here is a good blog post from my colleagues on one recent example from the Seventh Circuit. (Disclosure: Holland & Knight LLP, including the author of this blog post, represented a party in the identified case.)
We are paying special attention to matters where the alleged secrets are algorithms, model weights or training data. These disputes typically stress traditional "reasonable measures" and will continue to create interesting damages frameworks.