New York-based Pathos AI announced a $237 million Series E on October 23, 2026, alongside research partnerships with Bristol Myers Squibb and Tempus AI. The financing supports its AI platform and experimental medicines.
According to Pathos’s financing announcement, Bristol Myers Squibb led the round, which gave the company an approximately $3 billion post-money valuation. Existing investors also participated, although the release did not name them.
Reuters independently reported the funding and partnerships on October 23, identifying Pathos as New York-based. The announcement adds a substantial financing event to the city’s technology economy, with a business model that combines computational tools and the development of experimental medicines.
Two research directions, with different timelines
Pathos calls its AI platform Foundry. The company says it combines clinical, genomic and other information to guide drug-development decisions as evidence changes. The financing is intended to expand Foundry and advance its drug portfolio.
The separate collaboration announcement outlines a move into Alzheimer’s disease and other neurological conditions. Tempus is to assemble a de-identified dataset tracking patients over time, while Pathos plans to use Foundry to investigate disease biology and inform more targeted, adaptive clinical trials. Bristol Myers Squibb brings neuroscience drug-development expertise.
BMS also gains nonexclusive strategic access to Foundry across its broader portfolio and three-year exclusive rights to use the platform for its neuroscience development programs. That distinction matters: the announcement describes different access arrangements for broader research and neuroscience, rather than universal exclusivity over Pathos’s technology.
These are research and development plans. The October 23 announcements do not present new comparative clinical results demonstrating that the partnerships have improved treatment outcomes or increased the rate of drug approvals.
What the data agreement actually commits to
Tempus’s October 23 Form 8-K supplies details beyond the launch statements. It says the neurology order was signed on October 22. Tempus will use commercially reasonable efforts to deliver and license 500,000 de-identified patient records over five years, at 100,000 records annually. The records concern Alzheimer’s disease or other specified neurological conditions.
Pathos is to pay Tempus up to $55 million: $50 million in record-based license fees and $5 million for the availability of updates. The record fees depend on delivery. The planned dataset is therefore a multiyear undertaking, rather than 500,000 records already handed over at announcement. For researchers, that timetable makes data availability a milestone to track alongside the analyses the partners intend to perform.
The $55 million is a potential payment by Pathos to Tempus. It is separate from the $237 million equity financing raised by Pathos, and should not be added to the funding round.
The filing also discloses overlapping executive leadership and equity ownership between Tempus and Pathos. It says a special committee of disinterested Tempus directors reviewed and approved the neurology order and related side letter under the company’s related-person transaction policy.
Cancer programs remain a separate development effort
Pathos’s current pipeline page lists pocenbrodib, DO-2, P-100, previously called AZD4241, and JSKN016. The listed areas include prostate cancer, multiple myeloma, non-small-cell lung cancer, breast cancer and solid tumors. These programs should not be confused with a newly approved treatment arising from the neuroscience partnership.
The company explicitly identifies pocenbrodib as investigational and unapproved by the FDA or any other regulatory authority. Its description places the drug in a Phase 1 study for metastatic castration-resistant prostate cancer and a Phase 1b study for relapsed or refractory multiple myeloma, evaluated alone and in selected combinations. Its financing announcement describes advancing JSKN016 toward registrational studies as a priority; it does not establish that a new registrational trial has already begun.
There is a parallel oncology agreement between BMS and Tempus. In its own October 23 announcement, Tempus says BMS will license oncology data through 2031 and gain access to its Lens research platform. That arrangement is distinct from Pathos’s neurology data order and from the equity investment in Pathos.
For New York’s AI sector, the practical significance is the combination of software, research partnerships and a company-owned drug pipeline. Assessing its progress will require following both kinds of milestones: whether the data and analysis systems are delivered as planned, and what the experimental medicines show in clinical studies. A financing round can support those efforts; evidence of medical benefit has to come from the research itself.
