🌏 中文版
Funding Details
| Item | Value |
|---|---|
| Company | Pathos AI (New York City, USA) |
| Round | Series E |
| Amount | $237M |
| Lead | Bristol Myers Squibb (BMS) |
| Participants | Existing investors (undisclosed by name; prior backers include New Enterprise Associates, Revolution Growth, Lightbank, and Builders VC) |
| Valuation | $3B (up from $1.6B at Series D, a 1.9x increase in 17 months) |
| Total raised | $704M ($467M after Series D) |
| Founded | 2022 |
| Headcount | ~100 (Tracxn, September 2026) |
What this company does
Pathos AI is a clinical-stage biotech that develops its own drugs using AI — it doesn't sell AI tools to pharma companies, it uses its own AI platform to make portfolio and clinical-trial decisions and runs the drugs through the clinic itself.
Its core asset is Foundry, a platform built from thousands of specialized AI agents that continuously aggregate clinical data, genomics, biological-specimen data, and the shifting competitive landscape to decide what a drug program should do next. Unlike Isomorphic Labs or Recursion, which focus on using AI to predict molecular structures, Pathos differentiates at the decision layer: Foundry decides which clinical assets to in-license, how to adjust trial design, and which biological subpopulation to bet on — turning "making the right drug-development decision" into something that can scale.
Its clinical-stage pipeline currently holds four assets: JSKN-016, a TROP2/HER3 bispecific antibody-drug conjugate heading toward registrational studies; pocenbrodib, licensed from Novo Nordisk for metastatic castration-resistant prostate cancer and relapsed/refractory multiple myeloma; AZD4241, an ER PROTAC partnered with AstraZeneca; and DO-2, a next-generation MET kinase inhibitor. Its "customers" aren't subscribers — they're multi-year strategic partnerships with pharma and data partners like BMS, Novo Nordisk, AstraZeneca, and Tempus.
What this round signals
What it means for the agent ecosystem
Foundry runs thousands of AI agents that aggregate clinical, genomic, and biological-specimen data alongside the competitive landscape in real time — a rare case of agent architecture deployed at scale on high-stakes, long-cycle drug-development decisions, not the low-risk, high-frequency tasks typical of coding or customer support. The new capital will scale Foundry's capabilities and push JSKN-016 toward registrational trials, effectively using the raise to directly test whether "agents can make clinical decisions" holds up.
What the investors are betting on
The most important detail in this round isn't the dollar figure — it's the deal structure. In leading the round, BMS gets non-exclusive access to Pathos Foundry across its broader R&D portfolio, plus a three-year exclusive in neuroscience. This isn't a purely financial investment; it's equity bundled with platform access. At the same time, Tempus AI — co-founded by Pathos co-founder Eric Lefkofsky — signed a separate five-year, $55M data-licensing agreement with Pathos, under which Tempus builds a de-identified, longitudinal dataset on Alzheimer's disease for Foundry to analyze. In other words, BMS is betting on replicating Pathos's oncology decision-making playbook in neuroscience, not simply wagering that one AI biotech will get a new drug to market.
Numbers worth tracking
Conversion: baseline y=300, 60 units per billion dollars (y = 300 − value × 60), axis starts at 0. Data from Pathos AI's own announcements.
- Valuation climbed from $600M at Series C (October 2024) to $1.6B at Series D (May 2025, 2.7x in seven months) to $3B at Series E (October 2026, another 1.9x in 17 months) — no sign of slowing between rounds
- Total funding has reached roughly $704M behind a pipeline of just four clinical-stage assets — more than $175M in funding per clinical asset, far denser than a typical oncology biotech at the same stage
- The Tempus data-licensing deal has its own explicit price tag: five years, $55M. That money isn't being burned in the lab — it's buying data infrastructure for an entirely new disease area (neuroscience)
Watchlist status
Pathos AI is not yet on the watchlist. Recommend adding it under section D5 (Healthcare AI), tracking the Foundry agentic drug-development platform, this $237M Series E and BMS's equity-for-platform-access structure, and whether Pathos can actually replicate its oncology decision-making methodology in neuroscience.
Today's takeaway
I used to assume agent architecture mostly lived in low-stakes, retry-friendly tasks like coding or customer support. Pathos is running thousands of agents on drug-development decisions where a single bad trial design can burn a year — and a major pharma company is already backing that bet with real money. Just as notable is how BMS structured its investment: not a plain check, but equity bundled with platform access. That could become the template for how big pharma works with AI biotech going forward — investment not just as a financial return, but as a shortcut to acquiring in-house agent capability.
References
- Pathos AI Secures $237 Million in Series E Financing to Scale Agentic AI Drug Development Platform and Advance Oncology Pipeline
- AI drug firm Pathos raises $237 million, pushing valuation to $3 billion
- BMS leads Pathos AI's $237M series E alongside Alzheimer's link-up
- Pathos AI secures $365M series D to fund trial of Novo Nordisk solid tumor drug
- Pathos AI Closes $62M Oversubscribed Series C Round of Financing to Accelerate its Platform Approach to Drug Development
- Pathos - 2026 Company Profile, Team, Funding & Competitors