Insights

In its precedential July 2, 2026 decision in TrackTime, LLC v. Amazon.com Services LLC, the U.S. Court of Appeals for the Federal Circuit vacated a Delaware district court ruling that had found certain “executable program code” limitations indefinite and remanded for further proceedings. The Federal Circuit did not hold that such code language necessarily supplies structure or automatically avoids 35 U.S.C. § 112(f). Instead, it faulted the lower court’s analysis for treating the issue too mechanically and directed a more contextual inquiry into the full claim language, the recited function, and how a person of ordinary skill in the art would understand the disputed terminology.

For software and data-processing patents, the practical message is that §112(f) and indefiniteness disputes may turn heavily on the evidentiary record rather than on labels such as “code” or “logic” alone. Expert testimony and evidence showing whether relevant practitioners would recognize the term as naming a sufficiently definite structure can be important. Patent drafters should therefore avoid assuming that “executable program code” is a safe harbor; describing the operations performed, their technical relationships, and identifiable implementations more concretely may reduce later claim-construction risk and strengthen the record if validity is challenged.

The U.S. Congress is advancing the NO FAKES Act of 2026 as a federal response to AI-generated digital replicas and deepfakes involving a person’s voice or visual likeness. As of 8 August 2026, the latest formal action on the Senate bill, S.4591, remains its 24 June report from the Senate Judiciary Committee and placement on Senate Calendar No. 446; it has not become law. The reported text creates a transferable federal property right in voice and visual likeness, with post-mortem protection that begins at 10 years and may be renewed in five-year increments where qualifying authorised public commercial use continues, subject to a 70-year maximum. Recent industry analysis has focused on the bill’s platform-liability architecture: after receiving a compliant notice, an online service would address allegedly unauthorised digital replicas, while the uploader could file a counter-notification asserting that the material is not a digital replica, is authorised, or falls within a statutory exclusion. If the service elects to restore the material, restoration would generally occur no sooner than 14 days after the counter-notice unless the claimant has filed suit.

The structure borrows from the DMCA without simply copying it. Counter-notices would carry formal requirements, including a physical signature that is notarised or witnessed, contact information, a good-faith statement and consent to federal court jurisdiction; false notices or counter-notices may also create substantial liability. The bill separately carves out qualifying nonprofit libraries, archives, nonprofit educational institutions and related personnel, while preserving room for news reporting, commentary, criticism, scholarship and satire. The policy trade-off is practical rather than abstract: deepfakes can spread fast enough to justify rapid removal tools, but a one-way notice system can also suppress lawful expression. For platforms, media companies and AI providers, the harder compliance work may therefore lie in evidence handling, notice review, counter-notice workflows and litigation escalation rather than in takedown alone.

Mexico’s Institute of Industrial Property (IMPI) further clarified on 5 August 2026 how the new Technology Transfer Registry will operate as Articles 193–196 of the revised Regulations to the Federal Law for the Protection of Industrial Property are implemented. The registry is intended to cover technology-transfer arrangements, intellectual-property licences and franchise agreements, bringing filing details, agreement information and subsequent updates into a more defined compliance process.

For multinational companies, the practical issue is not the filing step alone but how registration interacts with enforceability against third parties, local tax treatment and later IP enforcement. Legal, tax and commercial teams should therefore align the licensed rights, contracting entities, payment terms and supporting documents before execution, while existing agreements may merit a review for any registration or updating requirement. A procedural omission can become disproportionately important when a licence later has to support a deduction, a defence or an enforcement action in Mexico.

The UK Intellectual Property Office (UKIPO) moved ahead with phase two of its One IPO digital patent service on 6 August 2026. The new stage pushes patent filing further toward structured electronic data and away from paper and conventional PDF forms. It also introduces dedicated metadata for AI-assisted drafting and inventorship, bringing the use of AI tools into the data checks performed during formalities examination.

The practical effect is clearest for applicants filing at scale through APIs or automated workflows. If a submission package omits the required AI-use declaration or contains incorrectly formatted fields, it may trigger a deficiency notice and affect eligibility for accelerated examination. The immediate issue is therefore not whether AI-generated material is patentable, but whether the filing can pass machine validation. Applicants using automation should treat AI-use records, inventorship confirmation and field mapping as part of pre-filing quality control rather than as an afterthought.

On 4 August 2026, the Munich central division of the Unified Patent Court (UPC) issued a procedural order in a dispute involving a communications standard-essential patent (SEP). With an invalidity counterclaim pending before the UPC and a parallel opposition proceeding before the European Patent Office (EPO), the court refused a request to stay the UPC case until the EPO process had run its course and instead placed the validity dispute on a tighter timetable.

The practical message is difficult to miss: a parallel EPO opposition may not provide a dependable route for slowing UPC litigation, and parties will have less room to hold expert evidence back until late in the proceedings. In SEP and other high-value European patent disputes, invalidity theories, technical expert positions and supporting documents may need to be developed earlier than some litigants are used to. Whether the same approach becomes consistent across UPC divisions remains to be seen, but the order reinforces a procedural culture that favours speed and disciplined evidence management over tactical delay.

WIPO’s International Bureau issued a Hague System practice update on 7 August 2026 confirming that multiple international design applications designating the European Union are no longer constrained by the requirement that all designs belong to the same Locarno class. As the EU design reform is implemented, applicants can therefore place designs from different classes—such as a product casing, a graphical user interface and packaging—in a single Hague application designating the EU, subject to the revised multiple-application fee structure.

The procedural change makes portfolio filing noticeably more flexible and may alter the cost comparison between the Hague route and direct EU filings. Consumer-electronics and cross-border brands can now consider grouping designs around a product launch or commercial portfolio rather than splitting them first by Locarno class. That flexibility should still be used selectively: combining designs can simplify filing and reduce fees, but disclosure timing, portfolio structure and the consequences of later disputes remain sensible reasons to separate some designs.