Pathos AI Raises $237M Series E for Drug Development

Pathos AI's expansion into neurological drug development brings a new data bill alongside its new capital. A five-year arrangement with Tempus AI gives the company a route to neurological patient records, with payments tied to delivery. The information needed to develop an AI business has an operating cost of its own.

The New York City company announced a $237M Series E on October 23, 2026, led by Bristol Myers Squibb alongside existing investors. Pathos put its post-money valuation at approximately $3B. The financing supports its Foundry AI platform and oncology pipeline while a separately announced collaboration takes that technology into neuroscience.

For CEO Iker Huerga, those activities bring several development calendars into the same organization. Pathos has its own investigational therapies to advance, a shared software engine to improve and a pharmaceutical partner preparing to use that engine across research programs. The investment connects the technology to more decisions about where drug-development resources should go.

What the Series E will support

Pathos identifies JSKN016 as a priority for the new funding, with plans to move the TROP2/HER3 bispecific antibody-drug conjugate toward registrational studies. The company also names pocenbrodib, AZD4241 and DO-2 among the programs the capital will support. These are development commitments, and the announcement supplies no completed therapeutic outcome for the financing to claim.

The current pipeline helps separate the programs. Pocenbrodib is being studied in prostate cancer and multiple myeloma. DO-2 targets MET-altered cancers, while P-100, previously AZD4241, is an estrogen-receptor degrader being developed for breast cancer. JSKN016 targets TROP2 and HER3 in solid tumors. Their biological questions differ even as Pathos proposes to share a technology foundation across them.

That creates an allocation problem as well as a research agenda. A platform team can improve a shared capability while an individual program needs a trial decision, a laboratory experiment or a development milestone. Pathos's proposed advantage lies in moving information between those jobs quickly enough to influence the next commitment. The capital has to support the programs producing that information as well as the system interpreting it.

How Foundry fits the operating model

Pathos describes Scout, Sprint and Foundry as connected parts of its development process. Scout evaluates potential assets and patient groups. Sprint teams take responsibility for moving individual programs toward their next milestones, while Foundry supplies a shared AI core connected to biological testing and clinical evidence.

Small program teams give that architecture an organizational consequence. The people working on a particular drug retain a focused development assignment, while AI scientists and engineers participate in the work surrounding it. A useful result from one program can inform the platform that supports another. That is the company's operating design, rather than a published measurement of how much clinical success improves.

The distinction matters when comparing AI businesses in healthcare. DevCuration's coverage of BioStack's medical-AI data work follows the construction of training and evaluation material. Pathos carries the additional responsibility of advancing its own drug portfolio. It is building software in an organization that must also make decisions about specific investigational medicines, giving the people developing the model a direct relationship with the work it is meant to support.

BMS becomes an investor and a platform user

Bristol Myers Squibb's collaboration with Pathos adds a commercial relationship to the financing. BMS gains non-exclusive access to Foundry across its broader portfolio and three-year exclusive rights for neuroscience. The collaborators intend to begin with Alzheimer's disease and other neurological conditions.

Those rights draw a boundary around how the technology can travel. Broad access allows a major drugmaker to put Foundry to work across research questions, while a defined exclusive field gives neuroscience a different commercial arrangement. For an operator evaluating a platform supplier, that distinction belongs alongside the technical description: availability can vary by therapeutic area even when the underlying engine is shared.

For Pathos, another organization's use introduces its own priorities and development context. The company will be serving a partner while continuing to develop its internal oncology assets. BMS has taken both an investment position and a place in the platform's intended use, making the relationship relevant to how Foundry develops beyond Pathos's own portfolio. Neither party has disclosed enough commercial detail to calculate the value of that access from the funding headline.

The data agreement has its own economics

Tempus's October 23 filing gives the neurology arrangement more precision than the partnership headline. Tempus will use commercially reasonable efforts to deliver 500,000 de-identified patient records over five years. Pathos's fees can reach $55M, with the data-license component dependent on records actually delivered and a separate amount for update availability.

That payment obligation is separate from the $237M equity raise. It describes resources Pathos expects to buy as it broadens the work its platform can support. Tempus also disclosed that a special committee of disinterested directors approved the relevant arrangements under its related-person transaction policy because of overlapping leadership and ownership. Those disclosed relationships are part of understanding how this ecosystem does business.

Pathos was founded by Eric Lefkofsky and Ryan Fukushima, according to its official company history, and Huerga became CEO in 2025. Its $365M Series D in May 2025 already supported an AI-and-oncology development ambition. The latest round extends that work into a larger set of obligations. As neurological records arrive and BMS brings new research questions to Foundry, Huerga's team will still have oncology programs moving toward their own clinical decisions.