Enlyte published its Q2 2026 nuclear verdict tracker on July 28, 2026, documenting four verdicts that together cleared $3.5 billion: a $1.6 billion wrongful death award against Upton Assets in Texas following a worksite explosion, a $603 million product liability verdict against a Chinese airbag manufacturer in Florida ($360 million punitive, $243 million compensatory), an $885 million pharmaceutical antitrust judgment in Massachusetts (Direct Purchasers v. Takeda Pharmaceutical), and a $420 million false advertising award against Papaya Gaming in New York (Skillz v. Papaya Gaming, the largest Lanham Act verdict on record). The report also documents state-level social inflation at 18% in Texas and 17% in Florida, both outpacing every state that passed tort reform in 2023 or 2024.
Who it affects
Self-insured employers with GL or commercial auto exposure in Texas and Florida carry the most immediate exposure. But the Q2 2026 dataset confirms that nuclear-scale outcomes are no longer confined to transportation or product liability defendants. The Massachusetts and New York verdicts reached pharmaceutical antitrust and commercial fraud, two lines that rarely appear in a transportation-focused risk manager’s reserve discussion, at the same dollar magnitude as a catastrophic trucking loss. Any program holding per-occurrence SIRs above $1 million in GL or commercial auto lines should treat the Q2 data as direct reserve evidence, not background news.
Reserve mechanism
The central problem is that expected loss rates carry an embedded severity assumption. For most self-insured programs, that assumption was anchored to experience through 2021 or 2022, before the 2023-to-2026 verdict escalation compounded. A development triangle built on that experience reflects a severity distribution that looks stable in cumulative paid data but systematically understates ultimates: the largest claims settle late or reach trial years after occurrence, so the severity shock does not surface in the triangle until it is already too late to catch an annual reserve study.
The RPS Q2 2026 Umbrella and Excess Market Update puts current casualty loss trends at 12-15% annually, driven by rising indemnity and defense costs, broader liability definitions, and juror desensitization to large awards. A program whose expected loss rate has not moved in three years is likely a full trend generation behind.
Enlyte’s report also flags AI-assisted case selection as a quantifiable accelerant: plaintiff firms are now using AI tools to identify and target high-potential cases faster, compressing the timeline from incident to outsized verdict. That selection effect does not appear in a standard frequency-severity triangle; it registers instead as an upward shock in severity development in the most recent accident years, precisely where the data is thinnest.
Florida’s tort reforms present a particular diagnostic challenge. The 2023 legislative changes dampened claim frequency, but Enlyte’s data shows severity still running at 17%. A program that reads the frequency improvement as evidence of overall loss improvement may be taking credit that the severity side has already given back.
Earlier geographic analysis based on 2024 data showed five states accounting for 76% of all nuclear verdict dollars; the Q2 2026 trajectory suggests Texas and Florida are both running above their prior-year shares. See Five States Hold 76% of 2024’s $31.3B in Nuclear Verdicts and GL and Commercial Auto Remain Above 100% CR Through 2026 for the combined-ratio context that frames why carriers are still unable to price these lines to profitability.
For programs carrying commercial auto exposure at a trucking or fleet operation, the severity trend mechanics are the same ones addressed in Commercial Auto and Fleet IBNR: the tail of large-verdict development in recent accident years is the principal source of reserve uncertainty.
What this means for your next review
Ask whether the expected loss rate in your GL or commercial auto reserve estimate was set using data through 2022 or earlier. If so, the Q2 2026 verdict distribution is evidence for a severity trend load, not merely context for a qualitative discussion. Also ask how the tail factor accounts for late-settling large claims: if it reflects historical development from a period before the current verdict environment, it is likely understating the tail.
Enlyte’s Q3 2026 nuclear verdict tracker is expected in October 2026. Any appellate reduction of the $1.6 billion Texas award will offer a read on whether nuclear verdicts are surviving post-trial review at the same rate as prior years; that outcome matters for whether the Q2 amounts represent true reserve exposure or headline numbers that courts will ultimately trim.