The 2026 Mass Tort Map

1. The shape of the docket

Read in aggregate, the federal mass tort system looks enormous and sprawling. Read by where the cases actually sit, it looks like a short list of mega-dockets with a long tail trailing behind them. As of the JPML's October 1, 2026 report, the system held 204,977 pending actions across 162 active multidistrict litigations, down from 206,182 across 166 a month earlier. The distribution is so lopsided that the headline count is almost the least useful number in it.

Dockets carrying 1,000 or more pending actions make up only about an eighth of all active MDLs, yet they hold the overwhelming majority of the cases. The concentration is not a quirk of one big year; it is the structural fact that shapes leadership appointments, settlement leverage, and where the plaintiff bar spends its advertising dollars.

Within that top tier, one docket dominates everything else. Johnson & Johnson's talc litigation alone accounts for roughly a third of all pending federal MDL actions, more than the next several dockets combined, which is why a single bankruptcy ruling in Texas can move the entire system's case count.

2. The growth is inside the mega-dockets

The long arc for active MDLs is still flat to slightly declining, down from roughly 170 in early 2025 to 159 this spring. The count climbed back to 166 by September and then gave up four dockets in the October report, which is the first month this cycle in which the total pending-action count fell as well.

What grows is the size of the dockets that already dominate. Talc is the clearest illustration: the count climbed steeply through mid-2025, plateaued near 67,000–68,000 after the proposed bankruptcy exit ramp collapsed, then resumed climbing at roughly 400 cases a month, adding 406, then 479, then 336 across the July, August and September reports. October added 90. One month is not a trend, and the number to watch is whether that slowdown holds.

J&J proposed a $5.5 billion resolution on July 27 covering roughly 76,000 ovarian cancer claims across federal and state courts, conditioned on 95% claimant participation, with the first payment capped at $3 billion in 2027. A deal that needs 95% of a pool to sign is priced against a pool that kept growing for three reports after it was announced.

The opposite shape, near-vertical growth that ran straight into a settlement, belongs to Depo-Provera. The docket added roughly 1,700 cases in a single report this spring, the largest one-month gain of any MDL this cycle, and on June 15 Pfizer and plaintiffs’ leadership disclosed a global settlement in principle: terms undisclosed, eligibility limited to cases pending in the MDL, and the first bellwether vacated. Filings did not stop. Another 322 arrived in the July report, 464 in August and 109 in September. The October report added 9, which is the first month the docket has effectively stopped taking on cases.

For anyone tracking where intake advertising and referral volume are flowing, this is no longer a growth story; it is an eligibility-window story, and the window is now the whole game.

What makes this docket worth watching is that the settlement and the science are being built at the same time. Pfizer and plaintiffs’ leadership signed a master settlement agreement on July 21, with terms confidential and the parties estimating that roughly 80% of MDL plaintiffs qualify. Registration is expected to run through November, with payments projected to begin in the first quarter of 2027.

Judge Rodgers appointed a settlement administrator on August 5. On September 4 she issued three orders in one day, naming an allocation master for the supplemental compensation fund, a qualified settlement fund administrator, and a lien resolution administrator. An administrative docket dedicated to the settlement followed on September 15.

Alongside that machinery sits a fully briefed Rule 702 record. Plaintiffs moved on August 5 to exclude five defense experts across the pilot cases. Pfizer moved the next day to exclude the plaintiffs’ general causation experts. Briefing closed on August 7 and the general-causation hearing was set for September 18. No ruling has issued. Judge Rodgers has status conferences set for October 16, November 20 and December 18.

That parallel posture is the thing to read. A settlement in principle usually slows the merits track; here the court issued a case management order for ongoing litigation five days after appointing the settlement administrator, and conditional transfer orders are still bringing new cases in. A defense ruling on the September record would leave the deal intact while stripping the value out of every claim sitting outside it, which is precisely the exposure the eligibility window was drawn to contain. The October report, with 9 new cases, suggests the filing bar read that risk the same way.

Where the growth actually went this month

Depo-Provera held the growth title for most of the cycle. The October report hands it to the social-media docket, which took in 616 new actions in a single month. Bard PowerPort added 406 and Uber passenger assault added 328. Those three dockets account for most of the month's new filings.

Read the percentage column separately from the count column. Acetaminophen nearly doubled on a base of 58 actions, which is a signal about one small docket rather than about the system. The platform and rideshare dockets are the ones moving real volume. All three sit in the Northern District of California, before Judges Gonzalez Rogers, Breyer and Lin.

The last row explains the headline. System-wide pending actions fell by 1,205 between the two reports, and Roundup alone shed 2,821. Strip that one docket out and the rest of the system grew by roughly 1,600. One docket emptied faster than the others filled, which is a different fact from a shrinking system. The JPML report states the count and not the reason, so confirm the mechanism in the transferee court's docket before building an argument on it.

3. The bellwether turn

For most of the last three years, the mega-dockets generated motion practice, not verdicts. That changed in the first half of 2026. Within roughly six months, the litigation produced the first jury verdict in the social-media docket, the first bellwether result in Bard PowerPort, a reset of the AFFF personal-injury track, and a string of state-court talc rulings. That is the moment a docket stops being a filing statistic and starts pricing risk. The second half of the year added the other half of the lesson: three of the largest dockets reached for settlement machinery before a second jury could be seated.

Social media: from "untested" to a plaintiff verdict

For three years the social-media docket was a Section 230 argument with no money changing hands. In early 2026 the lead California state-court bellwether broke that pattern: TikTok and Snap settled confidentially before trial, and on March 24 a jury found Meta and Google/YouTube liable for negligently designing platform features that harm adolescent users. The verdict followed a ruling that neither Section 230 nor the First Amendment barred the design-defect theory.

The first federal MDL bellwether never reached a jury: the Breathitt County school-district case settled with all four platforms in May, for a reported figure near $27 million, weeks before its June 15 trial date. Meta then resolved the state attorneys general cases on August 26, agreeing to pay 47 states up to $17.1 billion over ten years and to change teen-safety defaults on Instagram and Facebook. That money goes to state programs, and the consent judgment preserves the personal-injury and school-district claims.

The docket grew by 616 cases in the October report alone, to 3,824. A consolidated state-court bellwether is set for October 28 in Los Angeles, and the first federal school-district trial, Tucson Unified, is set for February 8, 2027.

Bard PowerPort: a "defense verdict" that wasn't clean

The first federal PowerPort bellwether went to a jury in Arizona and produced a partial defense verdict: the manufacturer prevailed on failure-to-warn, failure-to-instruct, and consumer-fraud claims, but the jury hung on the central design-defect question. Calling it a defense win overstates it. Additional bellwethers run through early 2027, and the court has pushed the parties into settlement discussions in July.

AFFF: water resolved, personal injury reset

The PFAS story has split cleanly. The public-water track is largely resolved through large settlements with the principal chemical defendants. The personal-injury track, organized around a defined set of qualifying conditions with kidney cancer as the lead, saw its first bellwether pulled off the October calendar and reset, even as the inventory of unfiled claims runs into the tens of thousands.

4. Two structural defenses under pressure

The throughline of 2026 is not any single verdict. It is that two defenses the corporate bar has relied on for years showed cracks in the same window.

Plaintiffs stopped litigating what users posted and started litigating how the product was built. That reframing is what carried a verdict past Section 230.

Section 230. The platform-immunity statute was built for liability over third-party content. The plaintiff bar's move was to stop suing over content and start suing over design: algorithms, notification cadence, infinite scroll, the architecture of engagement. Courts have increasingly let that theory through, and a jury has now accepted it.

The Ninth Circuit then narrowed what the statute does: on August 10 it dismissed the platforms’ interlocutory appeal as premature, holding that Section 230 supplies a defense to liability rather than an immunity from suit. A defendant that must try the case first and raise the statute on appeal from a final judgment is a defendant that carries trial risk on every file. The design-defect lane around the statute is now a proven path rather than a hopeful one.

Bankruptcy. The other shield is the use of Chapter 11 to corral and cap mass-tort liability. J&J's third attempt, the "Texas two-step" via a purpose-built subsidiary, was dismissed, with the court pointing to irregularities in how claimant support was tallied. The proposed multibillion-dollar resolution evaporated, and roughly 90,000 claims returned to active litigation. The lesson for any defendant eyeing the divisional-merger route is that the maneuver is getting harder to land.

What replaced it says as much: on July 27 J&J proposed a $5.5 billion resolution in the tort system, conditioned on 95% participation and excluding mesothelioma and future claims.

The company is also pressing an evidentiary attack in parallel. Plaintiffs withdrew two specific-causation experts in May after a special master faulted their methodology, and on July 22 Magistrate Judge Singh ordered plaintiffs to show why the remaining talc claims should not be dismissed for failure to prove specific causation. A defendant that loses the bankruptcy shield can still litigate the proof.

5. What the JPML centralizes, and what it refuses

The Panel's gatekeeping in this cycle rewards two things: a clearly defined injury and a recognizable, bounded set of defendants. Where either is missing, centralization tends to fail. The sharpest illustration came weeks apart, around the same product.

The contrast in the first two rows is the whole lesson. A theory built on a defined harm and a single defendant was centralized; a broader "addiction" theory that would have swept in multiple publishers and an open-ended product universe was refused. The same logic explains the GLP-1 split into two MDLs by injury type, and it is the single best predictor of which emerging theory becomes a docket and which stays a press release.

The GLP-1 dockets are where that split is now being stress-tested, and the numbers show why the division mattered. The gastrointestinal docket (MDL 3094) held 4,025 pending actions in the October report, having added 3 cases after months of growing at roughly 80 to 95; the NAION docket (MDL 3163) held 247, from 29 when it was created at the start of the year, and added 31 in October alone. Neither is close to the top ten. Both sit before Judge Marston in the Eastern District of Pennsylvania, which means the same court is running two causation records built on different science.

The gatekeeping inside MDL 3094 has been more consequential than its size suggests. In a memorandum order entered August 15, 2025, the court resolved the first cross-cutting issue against the plaintiffs: it excluded two of their experts on whether gastroparesis can be diagnosed without objective testing, declined to exclude the defense expert, and held that any plaintiff claiming gastroparesis must show a diagnosis based on a properly performed gastric emptying study, by scintigraphy, breath test, or wireless motility capsule.

Plaintiffs’ leadership has represented that the large majority of the docket pleads gastroparesis, so a single evidentiary ruling reset the qualification standard for most of the inventory. Verified claims and unverified claims stopped being worth the same thing on that date.

The second and third cross-cutting issues, warning adequacy with preemption and general causation, went to a five-day Rule 702 hearing that ran September 14 through 18, 2026. No ruling has issued. Judge Marston has status conferences set for October 13, November 17 and December 15, and no bellwether trial schedule exists.

The asymmetry is worth naming: a ruling for the plaintiffs moves 4,025 cases toward bellwethers, while a ruling the other way leaves a docket that grew every month for two years with no admissible path to a jury. The October report is the first in which filings into MDL 3094 nearly stopped, which is what a docket waiting on a causation ruling looks like.

The newest dockets on the board

The Panel centralized nine MDLs between June and August 2026. Six of the nine are commercial: two data breach dockets, two antitrust dockets, one patent docket and one securities docket. Three carry personal injury claims, and those are the ones a plaintiff-side firm reads first: Dupixent, the Boston Scientific spinal cord stimulator, and firefighter turnout gear made without AFFF.

Two of the new product dockets are already growing. Boston Scientific went from 29 pending actions on September 1 to 41 on October 1, and the turnout gear docket went from 8 to 10. Neither docket is large enough to price yet. Both were centralized in early June, and the change shown here covers one month of that run.

The October report carries no MDL centralized after August 18, so the Panel added nothing for roughly six weeks before the report date. Petitions pending at the Panel's fall hearing session are the place to watch for the next entries.

Ecosystem defendants are now the norm

The defendants in these dockets are rarely a single manufacturer anymore. PFAS names the chemical, formulation, manufacturing, distribution, and successor chain, reaching the principal chemical companies and their corporate descendants. GLP-1 names both Novo Nordisk and Eli Lilly across a drug class spanning eight branded products, while the court separates the claims by injury rather than by manufacturer. Talc has pulled consumer-product allegations into a litigation built on industrial asbestos. The practical drafting implication: defendant identity is a corporate-family question, not a product question.

6. What it means for plaintiff-side practice

Two operational shifts follow from the data. First, Rule 16.1, effective December 1, 2025, formalizes early case management: leadership structure, discovery planning, and early identification of the factual and legal issues. Expect courts and defendants to press sooner for proof of product use, exposure, diagnosis, and injury timing. The documentation burden moves forward in the case, which means intake quality, not intake volume, is where the risk now sits.

Second, the verdict turn changes how a docket should be read. A bellwether outcome is a data point about a single plaintiff, not a verdict on the inventory. The PowerPort result, prevailing on warnings while hanging on design defect, is the textbook example. Treating one bellwether as dispositive, in either direction, is the most common analytical error of the cycle.

The strategic picture for the plaintiff bar is a barbell. At one end, the proven mega-dockets (talc, AFFF, hair relaxer) with known injury definitions and recognizable defendants, several of which now turn on an undecided evidentiary motion: the hair relaxer general-causation challenge has been fully briefed before Judge Rowland since June 16, expert discovery closes November 16, and no bellwether has a trial date. At the other, a fast-growing breakout (Depo-Provera) and a coherent new category of platform-harm-to-minors cases (social media, Roblox, rideshare assault) where the law is being made in real time. The thin middle is the speculative theory the JPML has shown it will refuse to centralize.

Sources & method

Case counts are drawn from the JPML's pending-dockets and monthly statistics reports, anchored to the October 1, 2026 report; counts are point-in-time and move monthly. Month-over-month changes compare the October 1 and September 1, 2026 editions of the pending-dockets-by-actions-pending report. Centralization dates come from the date-transferred column of the pending-dockets-by-MDL-number report of October 1, 2026. That column records the date of the Panel's transfer order. The same report carries a separate date-filed column, which records when the petition reached the Panel. Verdicts, settlements, and procedural milestones are compiled from court records and litigation reporting by Reuters, Law.com/ALM, NPR, CNBC, and established docket trackers. Settlement terms described as proposed or signed but not finalized are reported as such; confidential terms are not reconstructed here. The GLP-1 orders and hearing dates are taken from the docket in MDL 3094, No. 2:24-md-03094-KSM (E.D. Pa.). Figures should be re-verified against primary sources before republication. Nothing here is legal advice.

Citing these figures. The case counts come from public JPML reports. Use them in a client memo, a pitch deck, a newsletter or a post. Cite Lawsuit Intelligencer with a link to this page, so your reader can check which month the numbers come from. Reproducing a table as it is arranged here, one of the charts, or a passage of the analysis asks for the same credit and link. To run the piece in full, write to editor@lawsuitintelligencer.com.