California Supreme Court Rejects a “Duty to Innovate” for Drug Manufacturers: What the Decision Forecloses and What It Leaves Open
On August 3, 2026, the California Supreme Court held in Gilead Tenofovir Cases (Aug. 3, 2026, S283862) ___ Cal.5th ___ (slip opn.), that a drug manufacturer owes no duty of care to users of a non-defective drug when deciding whether and when to commercialize an allegedly safer alternative drug. The Court reached that conclusion on facts in which the alternative compound had not completed large-scale human testing or obtained FDA approval, and its reasoning turns heavily on that posture. The decision reverses the January 2024 Court of Appeal ruling that had allowed plaintiffs in a coordinated proceeding to pursue negligence claims against Gilead Sciences without alleging that the drug they took was defective, and it directs the trial court to enter summary judgment for Gilead on all causes of action. For life sciences companies and their investors, the significance runs beyond products liability defense. The Court of Appeal decision had introduced a liability theory with no clear limiting principle into lifecycle management, diligence, and disclosure. That theory is now foreclosed in California on the facts presented, and the majority’s reasoning creates substantial obstacles to reviving it on other facts.
What the Court Held
Justice Groban authored the opinion of the Court, joined by Justices Corrigan, Liu, Kruger, and Desautels (sitting by assignment). Chief Justice Guerrero concurred in the result only and not in the reasoning. Justice Evans dissented. The majority opinion therefore commanded five votes, and the judgment six.
The Court did not hold that a product defect is a required element of every claim against a manufacturer. It instead assumed for the sake of argument that manufacturers may owe a general duty of reasonable care apart from the duty to market non-defective products, and then applied the Rowland v. Christian factors to find an exception to that assumed duty.
On foreseeability, the Court reasoned that a manufacturer cannot determine a drug candidate’s safety and efficacy during early clinical testing, so it cannot reasonably foresee that delay will harm users of an existing product. The causal chain was attenuated because it depended on uncertain scientific outcomes and independent decisions by regulators, physicians, and patients. On the policy factors, the Court found moral blame not meaningfully implicated given the “morally neutral and socially valuable” reasons that may drive development decisions, warned that the proposed duty would distort research priorities and invite hindsight-based second-guessing of complex scientific judgments, and concluded that the burden on manufacturers would be substantial. The Court framed its holding as reflecting the limits of common law negligence as a tool for supervising drug development in a field already subject to extensive federal oversight.
The Majority’s Reasoning Extends Beyond the Facts
The majority did not treat the plaintiffs’ theory as a close question. It observed that neither the plaintiffs nor the dissent identified a single decision, in California or any other jurisdiction, recognizing a duty to develop and commercialize an alternative drug to replace a concededly non-defective one, and it noted that the dissent’s 29 pages cite no case finding liability on that theory. It characterized the theory as novel and as requiring fact finders to reconstruct years of research and development decisions made against evolving scientific evidence and uncertain outcomes. It stated that negligence law is “poorly suited to supervise forward-looking decisions grounded in uncertain, incomplete, and evolving clinical data through retrospective assessments of reasonableness.” Rejecting the dissent’s characterization of the decision as conferring sweeping immunity, the majority said that what it declined to do was recognize, “for the first time anywhere,” sweeping liability for injuries from a concededly non-defective drug because the manufacturer allegedly failed to make a different drug available sooner. On the dissent’s patent-exclusivity argument, the majority said the dissent cited no authority for the proposition and that the Court was aware of none.
That matters for how the holding is likely to be applied. Although the Court framed its conclusion as an exception to an assumed duty rather than as a categorical rule, most of its policy analysis is not tied to the developmental posture of the alternative compound. The concerns about distorting research priorities, about retrospective reasonableness review of scientific judgment, and about the absence of any supporting authority anywhere apply to the theory generally. A plaintiff seeking to replicate the Court of Appeal’s holding on different facts would need to overcome that reasoning, not merely distinguish the stage of development. Taken together with the four justices who wrote or joined writings sympathetic to a defect requirement, the practical outlook for the theory in California is poor.
What the Court Did Not Decide
The formal reservations are nonetheless worth noting, because they define where argument remains available.
First, the Court expressly declined to determine whether a drug manufacturer may ever be liable in tort for negligent conduct causing injuries from a non-defective drug. Chief Justice Guerrero would have held that a product defect is a required element under Jiminez and Merrill. Justice Kruger, joined by Justices Corrigan and Desautels, wrote separately to underscore that the majority’s threshold assumption was made for the sake of argument only and to record shared skepticism that Civil Code section 1714 overrides the defect requirement. Four justices thus wrote or joined writings expressing sympathy with the defect requirement. The question is formally open, but the reservation is more technical than practical: the majority’s stated reasons for finding an exception would apply to a non-defective-drug negligence claim of this kind, however the duty question was framed.
Second, the holding is anchored to the developmental posture of the alternative product. The Court framed its foreseeability analysis around an alternative that had “not yet undergone large-scale clinical testing in humans” and had not received FDA approval. A claim premised on delay in commercializing an alternative that had already generated positive pivotal data, or that was already approved, is not squarely resolved. That is the boundary most likely to be tested. A plaintiff pressing it would still have to contend with the majority’s broader objections to supervising development sequencing through negligence law, which do not depend on the stage of testing.
Third, the Court declined to decide whether manufacturers outside the pharmaceutical context owe a duty to commercialize allegedly safer products in place of concededly non-defective ones, while observing that accepting the theory could have far-reaching consequences across a broad range of industries. For medical devices, diagnostics, and platform companies, the question is unresolved. The majority’s reasoning supplies arguments by analogy, but not protection.
Where Exposure Remains
The residual exposure is not a revival of the rejected duty. It consists of conventional theories that were always available and that the Court took care not to disturb, across both product liability and adjacent non-tort claims including unfair competition, fraud, antitrust, and securities law. The practical point for deal lawyers is that conduct once framed as a failure to innovate can be repackaged under those headings, not that the rejected duty survives in some attenuated form.
On the products liability side, the Court reiterated the rule of Brown v. Superior Court that drug manufacturers are not strictly liable for design defects. It noted separately that they remain subject to claims for negligent design, manufacturing defects, and failure to warn of known or reasonably knowable risks. The Court also observed that a plaintiff may still try to establish negligent design by pointing to an available alternative design whose benefits outweighed the burden of adopting it. Evidence about a next-generation compound therefore does not drop out of a product liability case; within that framework it must be connected to a defect in the marketed product. Nothing in the decision imposes that defect linkage on the non-tort theories discussed below.
Outside product liability, the decision does not eliminate consumer protection claims or common law fraud theories. The Court noted that plaintiffs may be able to pursue relief under applicable consumer protection laws and, in appropriate circumstances, under common law fraud principles, including on theories of withholding material information or misleading the public. The Court expressed no view on the merits of any such claim.
That is the likely migration path. Justice Evans’ dissent catalogs patent evergreening, product hopping, patent thickets, and pay-for-delay, which are the framings under which the same sequencing conduct is litigated as antitrust and unfair competition claims rather than as tort. Lifecycle decisions can be expected to continue drawing scrutiny through those channels, and through securities claims where public statements about next-generation programs and franchise transitions are at issue.
Practical Consequences for Portfolio and Deal Work
The majority’s reasoning gives weight to resource allocation toward more pressing medical needs and to the genuine uncertainty of early clinical data. Ordinary course records that accurately capture the scientific and portfolio rationale for sequencing decisions, prepared in the regular course rather than in anticipation of a claim, remain relevant to how those decisions are later understood. That is so across the unfair competition, fraud, and antitrust theories that the decision leaves available as well as in the products liability setting.
For transactions, the decision is a reason to reassess positions taken since January 2024. Diligence on a franchise pairing a marketed product with a next-generation asset has had to account for a liability theory with no clear limiting principle and no reliable damages ceiling, and that has surfaced in representation and warranty scoping, indemnity and escrow sizing, representation and warranty insurance underwriting, contingent value right and earnout constructs, and litigation contingency disclosure in registration statements. Whether and how those positions should change will depend on the particular franchise, the jurisdictions in issue, and the claims actually asserted or threatened.
Issuers that added risk factor language addressing the “duty to innovate” theory after the Court of Appeal decision may wish to revisit that language in light of the decision. Whether disclosure should be revised, and how, depends on each issuer’s own facts and its obligations under the federal securities laws. The reserved questions, the possibility that the theory is pled in other jurisdictions, and the continued availability of negligent design, failure to warn, unfair competition, and fraud claims all bear on that assessment.Gibson Dunn served as counsel for amici curiae the National Association of Manufacturers, the Alliance for Automotive Innovation, the American Tort Reform Association, the American Coatings Association, the American Chemistry Council, the Medical Devices Manufacturers Association, and the Consumer Technology Association, in support of the prevailing party. For the appellate treatment of the decision, see Gibson Dunn’s August 3, 2026 client alert, California Supreme Court Holds That Drug Manufacturers Cannot Be Held Liable For Selling Nondefective Drugs On The Theory That They Should Have Made Safer Drugs Available Sooner. Gibson Dunn’s Life Sciences and Appellate and Constitutional Law teams are available to assist companies and investors in assessing the decision’s implications for lifecycle management, diligence, and disclosure.