Pathos AI Raises $237m Series E Led by BMS

Bristol Myers Squibb has led a $237m Series E financing in Pathos AI, the drugmaker said on Thursday, valuing the AI-driven drug developer at a post-money figure of “approximately $3 billion” and tying the cheque to a wider data-and-AI collaboration that starts with Alzheimer’s disease. The pharma giant is not just taking a stake: it is buying access to Pathos’s software and to a large patient dataset that a third company, Tempus AI, will build.

Strip away the branding and the deal is a bet on a method rather than a medicine. Pathos says its platform, Foundry, runs “thousands of specialized AI agents” that continuously synthesise clinical data, genomics and the competitive landscape to decide a drug programme’s next move. None of the therapies in play is approved, and the probability-of-success language throughout is Pathos’s own; what BMS is funding is the attempt.

In brief

Bristol Myers Squibb led a $237m Series E in Pathos AI at a post-money valuation of about $3bn, alongside existing investors.
The round comes bundled with a neurology collaboration: Tempus will build a de-identified dataset Pathos licenses for up to $55m over five years, and BMS gains access to Pathos’s Foundry platform.
Every therapeutic claim is a development plan, not a result; no drug is approved.

Bristol Myers Squibb is buying access, not just a stake

Alongside leading the raise, BMS secures “non-exclusive strategic access” to Foundry across its broader R&D portfolio and “three-year exclusive rights” to use it for neuroscience, the companies said. The pitch is that pairing BMS’s drug-development know-how with Pathos’s agents and a purpose-built dataset can speed decisions in Alzheimer’s, a field littered with costly failures. Pathos chief executive Iker Huerga framed it as combining “the best longitudinal data, AI that can reason over it at scale, and the drug development expertise to turn insight into medicines”. It is the kind of agentic-AI-meets-pharma tie-up TSN has tracked repeatedly, from OpenAI’s purchase of Orchestra Bio to insitro’s collaboration with AstraZeneca.

The Tempus data deal is worth up to $55m to Pathos’s supplier

The neurology data comes from Tempus, which will build a “de-identified, longitudinal dataset from hundreds of thousands of patients with neurological conditions”. Pathos will license it “under a five-year, $55 million agreement”. Tempus’s Form 8-K puts numbers on it: the Neuro Order Form, signed on 22 October, covers 500,000 de-identified records delivered at 100,000 a year, for fees of “up to $55 million” — $50m at $100 a record plus $5m for updates. A separate oncology statement of work with BMS adds $137.5m, so the two agreements “together provide for aggregate cash payments to Tempus of $192.5 million over approximately five years”. The filing also flags that, because Tempus and Pathos share leadership and ownership, the arrangement was treated as a related-person transaction and approved by “a special committee of disinterested directors” of Tempus’s board.

Four clinical assets, but no approvals and plenty of risk

The money is meant to push an oncology pipeline Pathos says Foundry assembled through “rapid strategic transactions”: four active clinical assets. The headline priority is JSKN016, which Pathos calls “a first-in-class TROP2/HER3 bispecific antibody-drug conjugate” that the funding will carry “toward registrational studies”; the others are pocenbrodib, AZD4241 — “a novel ER PROTAC partnered with AstraZeneca” — and DO-2, a MET kinase inhibitor. The caution worth stating plainly: “toward registrational studies” is not approval, and clinical-stage assets fail routinely, especially in neuroscience. The structural knot is that the data supplier, the AI developer and the lead investor are now financially entangled, which is why Tempus needed a disinterested-director sign-off. For readers, the near-term signal is not a new drug but whether a data-rich, AI-steered approach actually raises the hit rate — the same open question hanging over Deep Origin and BioAge’s lab-stage collaboration. UK readers have a stake too: both BMS and AstraZeneca run large UK operations, and Alzheimer’s weighs heavily on the NHS. The next concrete milestone Pathos names is the start of a registrational study for JSKN016.