Venue Realignment, PTAB Crosswinds and the Rise of AI Infrastructure Disputes

Texas continues to be the most consequential forum for patent disputes in the United States, but the shape and magnitude of that influence transformed throughout 2025 with three forces primed to drive outcomes in 2026. First, venue dynamics in the Western District of Texas and the Eastern District of Texas have realigned, solidifying the Eastern District’s position as the national volume leader with credible trial timelines and the Western District as another important venue for patent litigation. Second, the relationship between district courts and the Patent Trial and Appeal Board (PTAB) has been reshaped by recent changes to the discretionary review of petitions for inter partes review (IPRs), including proposed rules that would require sweeping stipulations from IPR petitioners. Third, a new wave of cases targeting artificial intelligence infrastructure and advanced semiconductor systems has landed in Texas courts, bringing renewed focus to subject matter eligibility and rigorous analysis of damages frameworks. This chapter of the guide synthesises these developments, and the message for practitioners is pragmatic:

• retire playbooks premised on divisional judge certainty;
• narrow in on invalidity strategies earlier; and
• front-load technical and damages work for platform-scale disputes that will move on a short fuse in Texas.

The Texas map has changed: venue realignment and case management

Western District of Texas after random assignment and the Austin shift

The Western District of Texas no longer revolves around a single, predictable patent docket in Waco, Texas. In 2022, the district instituted district-wide random assignment for cases filed in Waco, replacing a regime in which a Waco filing guaranteed assignment to the sole district judge sitting in that District, Judge Alan Albright. The practical meaning of that random assignment was straightforward but significant. After that general order, a complaint filed in Waco could be assigned to any of a number of judges across Austin, San Antonio, El Paso, Midland–Odessa, and other Western District divisions, with each judge running a distinct schedule and demonstrating different appetites for early motions and stays. The only sure assignments were matters “related” to prior Waco-based matters. In 2024, the district tightened its framework by requiring concrete, fact-based showings before cases are reassigned as “related”, closing a pathway that had allowed serial filers to steer matters back to their preferred docket. The net result is a district that remains experienced and potentially fast but no longer monolithic.

The centre of gravity has also shifted geographically. With approvals from his colleagues and the Fifth Circuit’s Judicial Council, Judge Alan Albright moved his chambers to the Austin Division in 2025 and began receiving a defined share of Austin’s civil and criminal docket while maintaining Waco assignments until that seat is filled. The shift has practical effects that sophisticated filers are already exploiting. Although random assignment remains, filing in Austin can increase the probability, without guaranteeing it, of assignment to Judge Albright. Indeed, 40% of the civil docket in the Austin Division will be automatically assigned to Judge Albright, with Judge Pitman and Judge Ezra also receiving a share of the Austin civil docket.

A further ripple has been the rise of divisions that offer predictability for reasons unrelated to Waco. Midland–Odessa, where a single Article III judge, Judge David Counts, presides, has attracted a non-trivial share of new filings as parties test whether a single-judge division can approximate the scheduling certainty they once sought in Waco. Although not surprising given frequent referrals to experienced Magistrate Judge Derek Gilliland, the Midland–Odessa division is not a one-for-one substitute for the Waco division. Convenience factors can look very different, and motions to transfer under 28 U.S.C. § 1404(a) could be a concern when a case’s factual ties to the division are potentially thin. Even so, the experiment underscores a broader point for practitioners: when assignment practices change in one judicial district, filings will migrate to other districts where parties perceive a potential for predictability and rapid case advancement.

For teams planning to file in the Western District, two operational adjustments deserve emphasis. Expectations for time-to-trial, stays and early motion practice need to be set division by division and, once assignment occurs, by judge. A complaint filed in Austin, San Antonio or Midland–Odessa may encounter different attitudes towards early motions under 35 U.S.C. §§ 101 or 112, different claim construction pacing, and different propensities to grant stays pending IPR activity or re-examination. Early motion practice, in particular venue and transfer, should be front-loaded and backed by concrete evidentiary support regarding witnesses, sources of proof and logistics.

The Eastern District’s resurgence and a reframed remedies conversation

About four hours north-east of Central Texas, the Eastern District of Texas has reclaimed its position as the busiest patent forum in the country. By 2024, roughly one in five patent cases nationwide was filed in East Texas, with one judge in the Marshall Division assigned nearly 800 new cases in 2024. The volume matters because it signals not simply filings, but jury activity, verdicts and a case-management culture that moves parties through claim construction and into trial on schedules credible enough to drive business decisions. National damages totals rose to more than USD4.3 billion in patent cases in 2024, a figure that reflects both the rebound in filings and a resurgence in jury trials, even though some headline-grabbing awards were later trimmed or reversed. For patent litigation practitioners, the practical takeaway is that Eastern District of Texas cases tend to set the cadence for entire patent portfolios. Anchoring a campaign in Marshall can influence the timing of parallel matters and shape the sequencing of any validity challenges in other forums.

Remedies in the Eastern District have also evolved in a way that, while not doctrinally transformative, are practically significant. Although injunctions have not become more common in the Eastern District, injunction motions are becoming more frequent and have become more credibly argued. The potential for injunctive relief can shape settlement dynamics and is something practitioners should account for in litigation, even when the operative goal appears to be only for leverage rather than ultimate injunctive relief. That planning requires technical teams to assess feasible design-arounds and economic teams to consider valuation frameworks earlier and on a more compressed timeline. Litigation posture should be aligned with product roadmaps so that design-around feasibility and timing are known variables not late surprises.

Transfer, consolidation and limitations on judge shopping

Policy and practice together have significantly reduced judge shopping in Texas patent cases. In 2024, the Judicial Conference announced guidance promoting random assignment in certain civil cases and discouraging practices that allow litigants to select particular judges by filing in single-judge divisions. While that guidance is not binding and is focused primarily on cases affecting government action, it crystallised a broader direction. Within the Western District, the 2024 order tightening related-case consolidation functionally closed the largest remaining lever for returning serial cases to a preferred judge. Combined with the 2022 randomisation in Waco and the ultimate judicial reshuffling in the Western District, the result is that divisional filing no longer guarantees assignment, and consolidation on one judge’s docket must be earned through concrete showings of overlap and economy.

At the same time, appellate decisions have sharpened the law of transfer. A series of Federal Circuit mandamus rulings beginning in 2020 emphasised that transfer motions are top priority, discouraged reliance on aggressive trial settings to defeat transfer, and required convenience factors to be grounded in specific facts about witnesses and sources of proof. More recent decisions in the Fifth Circuit have clarified that a movant must show the transferee forum is clearly more convenient in the aggregate and that any gains in convenience will be meaningful and likely to materialise. Courts have been cautioned against over-reliance on docket congestion, directed to assess local interests by reference to where relevant events occurred rather than where parties maintain offices, and encouraged to recognise that while electronic evidence reduces the weight of document location, the physical location of source code access or prototypes can still matter.

The takeaway on transfer is that motion practice will reward parties who build evidentiary records early. Movants should be prepared with declarations identifying specific likely trial witnesses, pinpointing where they live and work, and explaining why travel burdens differ across venues. They should document where physical evidence, prototypes and source code are accessed, who controls that access, and what trial logistics would realistically look like in each forum. Non-movants should meet that specificity with their own detailed showings, emphasising witnesses within subpoena range, Texas ties that are genuinely linked to the events giving rise to the case, and the practical realities of trying a case where the evidence and people are. Consolidation, while still prevalent in Texas cases, is not as much a sure thing. Plaintiffs should expect contested consolidation even when patents overlap, and defendants should resist consolidation by highlighting distinct accused products, different claim sets, and divergent functionalities. Of course, judicial economy remains a strong consideration in Texas, particularly where claim construction or summary judgment rulings already exist on the same patents.

PTAB–district court interplay under strain: stays, IPRs, re-examination and early elections

Director-led institution decisions and proposed rules narrowing IPR applicability

The familiar playbook of filing an IPR to challenge patentability and moving to stay district court litigation is under pressure in 2026 because of two developments. First, the Director of the Patent Office is personally deciding all institution decisions, with discretionary review formalised and bifurcated from a merits review. Previously, a single review was delegated to PTAB merits panels. Discretionary considerations tied to parallel litigation – such as trial schedule, overlap of issues, and the risk of duplicative proceedings – are now addressed within a Director-led framework that reflects a policy preference for a single forum rather than courts and the Patent Office acting in parallel.

Second, the Patent Office published proposed rules for comment in October 2025 that would materially restrict access to IPRs. The proposal would require petitioners to stipulate that, if an IPR is instituted, they and their privies will not pursue any invalidity grounds under 35 U.S.C. §§ 102 or 103 in any other forum. The proposed rules would also bar institution if validity had already been upheld in a prior forum, including district courts, the International Trade Commission, earlier PTAB trials, ex parte re-examinations prompted by third parties, or Federal Circuit decisions. The proposed rules would further deny institution when another proceeding is likely to reach a decision ahead of the PTAB timeline for its final written decision. Although there is an “extraordinary circumstances” exception, it is narrow by design and does not include learning of new prior art or expert analysis.

Although the proposed rules are not yet in force and are subject to change, the direction is clear enough that in-house counsel should plan for them now. Requiring petitioners to waive all other anticipation and obviousness defences as the price of institution will feel too costly in many product cases, especially where on-sale or public-use art is potentially available because the proposed rules would require that they be waived yet they still cannot be part of the IPR proceeding. Further, treating prior validity decisions across all forums as bars to institution will prevent serial challenges and highlight the importance of choosing the initial forum with care, if choices are even available. If finalised in 2026, these rules will force earlier elections between PTAB and district court, with less room for hedging, and they may increase the appeal of ex parte re-examination for defendants who want to avoid the proposed stipulation and estoppel.

The potential ex parte re-examination surge and its practical consequences

Ex parte re-examination (EPR) filings rose substantially in 2024 and even more in 2025, with a notable decrease in overlapping challenges in EPRs and IPRs. That pattern signals that defendants are likely already turning to EPRs when fast Texas schedules are making IPRs less likely to be instituted. An EPR offers different trade-offs than an IPR, including that patent owners can sometimes craft a record to create new prosecution history evidence to support claim constructions, potentially narrowing claims to avoid art without sacrificing infringement-friendly constructions. Courts also appear to be less likely to pause discovery, claim construction or trial settings on fast-moving Texas dockets because of EPRs. But by building early, case-specific, fact-rich records, demonstrating real overlap, and a credible path to simplification of issues, parties may seek a stay pending an EPR. Given that outcomes will turn on early facts, in-house counsel should treat the first six to eight weeks after service as decisive for developing a comprehensive validity strategy, including whether to pursue EPR, IPR, or neither, and how best to position any related stay motion.

Lower IPR institution rates are channelling filings to EDTX

The evolving IPR practice has altered the risk–reward calculus for suing and being sued in Texas. After the Patent Office centralised institution in the Director and floated rules that would require sweeping stipulations from petitioners, the IPR institution rate moved from historical norms around 67% to effectively zero for a period and now appears to be stabilising near the 30%–40% range. Plaintiffs increasingly view IPRs as a less certain and less immediate threat, which reduces the expected cost of defending a patent in parallel PTAB proceedings and shortens the time horizon to a merits decision in district court. That shift is showing up in filing behaviour. The Eastern District, already the national volume leader, is drawing a larger share of new cases as patentees seek fast and nearly certain trial timelines without the drag of PTAB detours.

The practical consequences of these changes are already coming into focus. Plaintiffs can model lower PTAB spend and less risk of delay, which supports increased filings and a push for earlier settlements anchored to district court schedules. Defendants should anticipate more validity litigation in court, with fewer opportunities to pause the case on the strength of a likely IPR and with greater reliance on EPRs where timing and estoppel considerations favour that path. Budgets should be reallocated accordingly: less for serial PTAB campaigns, more for early prior art development, claim construction and expert work in Texas. If the proposed IPR rules are finalised, the new equilibrium will harden, further reducing petitioners’ appetite to pursue IPRs and reinforcing Texas courts as a preferred forum.

What is being litigated now: AI infrastructure and semiconductor hardware

Why Texas?

A meaningful slice of the first wave of artificial intelligence infrastructure and advanced hardware cases has landed in Texas courts. These cases differ from the better-known AI copyright disputes. They target the backbone of modern AI and high-performance computing:

• interconnect technologies that link accelerators;
• data-processing units embedded in server architectures;
• memory hierarchies and scheduling systems;
• cluster orchestration software; and
• data centre networking used to train and serve AI models at scale.

Plaintiffs have filed cluster computing and accelerator patent suits in the Eastern District against major cloud providers, while the Western District has seen suits tied to data centre architectures and training pipelines. See, eg, ParTec AG & BF exaQC AG v Microsoft Corp., No. 2:24‑cv‑00433 (E.D. Tex.); Xockets, Inc. v NVIDIA Corp. et al., No. 6:24‑cv‑00453 (W.D. Tex.); Xockets, Inc. v Amazon.com, Inc. et al., No. 1:25‑cv‑01021 (W.D. Tex.).

Texas is a natural forum for these disputes for historical and practical reasons. Judges and juries in the Eastern and Western Districts have deep experience with complex semiconductor and networking cases. The dockets offer credible trial timelines that business decision-makers trust, which means that district court schedules can anchor broader resolution strategies across portfolios. The evolving assignment and transfer rules require more nuance in filing decisions, but the overall result is a forum where technically sophisticated disputes move promptly enough to drive business outcomes.

Evolving patent doctrines: subject matter eligibility and damages analysis

These AI infrastructure and semiconductor cases implicate two important patent doctrines: subject matter eligibility and damages. Subject matter eligibility under 35 U.S.C. § 101 remains a risk for software-oriented claims, particularly those that can be characterised as abstract data organisation or implemented with generic components. Claims that recite functional outcomes in broad terms – such as optimising task scheduling across a heterogeneous compute cluster – are especially vulnerable if they lack the kind of technical implementation detail that courts have recognised as transforming an abstract idea into a practical application.

Damages volatility has been evident in recent verdicts, and Texas courts have responded by scrutinising expert methodologies and the comparability of licences offered as benchmarks. Apportionment remains central to the damages inquiry. The value of the patented feature must be separated from the value of the unpatented system, and royalties must be tied to the incremental value contributed by the accused functionality. Courts in the Eastern and Western Districts have enforced these principles with increasing rigour, trimming or excluding damages models that rely on licences lacking genuine technological comparability or that fail to adequately apportion. That scrutiny has practical consequences: early damages assessment is essential. Sales and pricing data by feature, accurate mapping of accused features to revenue, and carefully developed narratives around how the patented functionality drives value should be identified early enough to withstand gatekeeping and cross-examination.

Remedies and design-arounds in platform-scale disputes

Remedies strategy in AI infrastructure and semiconductor cases requires integration with product planning from the outset. Although injunctions remain difficult for licensing-focused patentees to obtain, the reframed conversation on injunctions has made tailored injunction motions a more credible element of early leverage in certain cases. Plaintiffs increasingly propose relief constrained to specific product lines or features and structured to allow transition periods, while defendants respond by demonstrating feasible design-arounds and challenging causal nexus where the accused feature is not a primary driver of demand. These dynamics place a premium on early design-around assessments. Defendants should evaluate whether disabling or re-implementing accused features is feasible, how quickly alternatives could be deployed, and what trade-offs would result. These assessments can mitigate injunction exposure and shape settlement valuation.

On the flip side, plaintiffs should anticipate challenges to causal nexus and be ready with evidence that credibly ties demand for the accused products to the patented features. That evidence often requires connecting usage metrics, performance benefits and drivers of consumer demand to the claimed invention. Given that Texas courts typically move cases to claim construction and trial efficiently, the window for building these records is short. Practitioners should assume that technical and damages work will be needed on a short timeline and plan accordingly.

Conclusion

Texas remains the epicentre of US patent litigation, but it is a slightly different Texas than many corporate legal departments remember from just a few years ago, perhaps slightly more reminiscent of the pre-TC Heartland East Texas heyday. The Eastern District has re-emerged as the leading patent venue with steady case management, active juries and credible trial timelines. The Western District’s district-wide random assignment and its 2024 consolidation order ended the era when divisional filing could reliably control judge assignment, while Judge Albright’s move to Austin created a probabilistic path to his docket without restoring pre-2022 certainty. Transfer law has matured towards a clearer, stricter standard that rewards evidentiary detail and penalises generalities, and judge shopping has been replaced by careful venue analysis and disciplined motion practice. The PTAB–district court interplay is being rewritten by Director-led institution decisions, a noticeable rise in EPRs, and proposed rules that would require broad stipulations and restrict institution when other forums are poised to decide validity. If finalised, those rules will force earlier forum elections and reduce hedging between IPR and district court defences. Meanwhile, Texas is attracting next-generation disputes targeting AI infrastructure and semiconductor hardware, which test the boundaries of eligibility and damages frameworks.

Texas courts will continue to move cases efficiently. The parties that succeed in 2026 will be those that adapt to the new Texas venue realities, anticipate PTAB crosswinds, and meet platform-scale disputes with the technical and economic depth they require.