The 2026 Patent Strategy Reset: What Both Sides Must Do Differently

Patent litigation hit a ten-year high in 2025, and the Eastern District of Texas is again the busiest patent court in the country. At the same time, the Patent Trial and Appeal Board (PTAB) entered its sharpest slowdown in a decade. The familiar playbook of filing an inter partes review (IPR) petition and moving to stay the district court case is no longer reliable. Meanwhile, Section 101 eligibility questions, which decide whether an invention is the kind of subject matter that can be patented at all, hang over the AI patents now entering litigation. Both patent owners and accused infringers need to revisit assumptions they held three years ago.

Rethink intellectual property clearance and freedom-to-operate

With the Eastern District of Texas leading all federal courts in patent filings and Judge J. Rodney Gilstrap’s docket pushing cases to trial on aggressive timelines, companies cannot count on early exits. Nationally, plaintiffs filed 4,547 patent cases in 2025—the highest total in a decade—and nearly half of all non-practicing entity (NPE) district court activity in the first half of 2026 was in the Eastern District. For patent owners, the concentration of cases in East Texas is an opportunity. A firm trial date is leverage. It is also a deadline: whatever weaknesses a case has when it is filed are likely the weaknesses it takes to trial. So plaintiffs must pressure-test all key issues—infringement, invalidity, and damages—before filing, not after Markman, the hearing where the court construes the disputed claim terms.

For accused infringers, the venue question matters less than it once did. What matters is having a defense strategy that does not depend on delay. The two levers that used to supply it have weakened at the same time: IPR institution rates have fallen sharply, and East Texas trial dates are holding mostly steady. Stays are virtually non-existent. Freedom-to-operate analysis must account for the likelihood that any dispute could go the distance. Invest more in pre-launch clearance, identify design-arounds before products ship, and budget litigation as a cost of market entry, not an unlikely contingency.

Adjust validity strategy for the new PTAB reality

The biggest change of the last year is the slowdown at the PTAB. Parties filed only 201 PTAB petitions in the first half of 2026—fewer than in any half-year period in the last decade. The institution rate has fallen from 68% in fiscal year 2024 to 50% in fiscal year 2025, and monthly rates since have run as low as the 20s, with institution decisions moving from panels of administrative judges to the Director personally. Ex parte reexaminations picked up some of the slack for a while, but courts rarely stay a case for a reexam, and a requester has little role after filing—so there is less leverage to trade on.

Discretionary denial now dominates the front end of the process, and it is what most often keeps a challenge from being heard on the merits. If a parallel district court case is likely to reach validity first, that cuts hard against institution. A separate factor weighs against challenging patents in force six years or more. One factor runs the petitioner’s way: the USPTO has weighed whether the challenged patent threatens products made in the United States, a point challengers should document early.

But two caveats apply. Institution is no longer the end of the inquiry: the Director has vacated grants months after institution. And a rule proposed in October 2025 would bar institution when another forum is likely to reach validity first, and would require petitioners to give up their prior art defenses everywhere else.

Accused infringers must therefore build validity defenses for district court, not rely on the PTAB to do the work. Prior art searches should happen earlier and go deeper. Patent owners considering enforcement, meanwhile, should consider whether their patents would survive the new institution factors. A patent in force more than six years and covering foreign-made accused products has better odds of avoiding institution if a petition is filed, and parallel litigation makes institution less likely still. Prosecution strategies should account for these factors from the start, and companies should work closely with their portfolio management counsel so that filing, prosecution, and enforcement decisions account for these factors on the front end.

Prepare for Section 101 to decide AI cases

The first AI patent suits are targeting the hardware underneath the models: data processing units (DPUs), supercomputing architectures, and specialized chips. One suit over Microsoft’s AI supercomputing systems, ParTec AG v. Microsoft, was about a month from trial in the Eastern District before it settled in May. Hardware claims do not commonly face eligibility challenges, so the fights are typically about claim construction and damages apportionment. But when claims are directed at the AI models themselves rather than the machines they run on, Section 101 will decide cases.

The Federal Circuit has held that applying generic machine learning to a new problem is an abstract idea unless the patent improves the model itself. For patent owners, that means prosecution strategies for AI inventions must focus on technical improvements—not just novel applications. For accused infringers, eligibility challenges remain a powerful early defense against software-focused AI claims. Legislation to rewrite Section 101 got its first full committee hearing in July, but until Congress acts, the current framework will shape which AI patents are enforceable.

For companies developing AI products, drafting choices made during prosecution will determine whether a patent is valid at all. Patents that tie AI functionality to specific hardware implementations or that claim concrete technical improvements to model architecture are more likely to survive both eligibility challenges and PTAB review. Generic method claims that apply existing techniques to new data sets are exposed to both.

Conclusion

The old playbook—file an IPR, move to stay, and wait—no longer works. Patent owners must pressure-test their claims before asserting them and draft applications with the new PTAB factors in mind. Accused infringers must build validity defenses for district court and treat litigation as a likely outcome rather than an avoidable one. And everyone in the AI space must account for Section 101 realities (and ambiguities) that will not change until Congress acts. These adjustments are cheaper to make on the front end than after a complaint arrives.