Capital, shared data and new tools for AI drug discovery

1. Enveda announces $311 million to advance AI-discovered medicines¹

Enveda has announced the close of a $311 million Series E financing led by Catalio Capital Management. The company says the proceeds will move its investigational medicines into later-stage trials, bring additional inflammatory and metabolic programmes into clinical testing, and expand PRISM, its AI platform, alongside the automated laboratory supplying its data. The funding gives natural-product-based AI discovery a larger route into clinical development, where its predictions must face patient evidence. This is a financing announcement: the company’s early clinical claims do not establish the effectiveness of the platform or its unapproved medicines.¹

2. Four drugmakers join AISB Bind federated modelling initiative²

Apheris has announced AISB Bind, a collaboration with AbbVie, AstraZeneca, Bristol Myers Squibb and Johnson & Johnson to train molecular-binding models across proprietary pharmaceutical data. The initiative combines structures, affinity and potency measurements, and screening results through federated training; Apheris says each company’s underlying data stays within its own environment. Training is scheduled to begin this month, with final models expected in early 2027. The practical opportunity is broader industrial evidence for virtual screening and lead optimisation. Those are programme aims: the announcement does not yet demonstrate the new binding model’s prospective performance.²

3. Iambic files for a public listing as its oncology pipeline advances³

Iambic Therapeutics has filed an initial public offering registration statement and applied to list on Nasdaq under the symbol IAM. Its filing describes an AI-led discovery platform and an oncology pipeline led by IAM1363, a HER2 inhibitor in an ongoing Phase 1/1b trial. The company anticipates a registrational study as early as 2027, subject to regulatory feedback. The filing makes a consequential transition visible: computational discovery programmes are seeking public capital for clinical development. An S-1 is a proposed offering, not a completed financing, and the projected trial timetable remains the company’s plan.³

4. OpenADMET moves into a metabolism-focused second phase⁴

OpenADMET says its ARPA-H-supported AVOID-OME effort is moving into a second phase focused on drug metabolism. The consortium plans to screen 15 assay targets against tens of thousands of small molecules, expand structural coverage of metabolic enzymes and nuclear receptors, and continue quarterly blind challenges with open data releases. For model developers, the consequential change is a committed programme of prospective metabolism data and evaluation, beyond a single retrospective benchmark. These are announced next-phase activities, not completed datasets; the post does not establish that all planned assays or structures are already available.⁴

5. OpenADMET releases a tabular-model workflow after its PXR challenge⁵

OpenADMET has added a tabular-foundation-model workflow to its openadmet-models toolkit after examining approaches used in its PXR blind challenge. The implementation combines molecular embeddings with predictions from a cheaper primary screen, giving developers a reproducible configuration to explore learning across assay types. In its accompanying analysis, the team reports improved error on the examined split, but the implementation follows the completed challenge and is not a new blinded victory. The useful development is an accessible modelling workflow, with enough detail to inspect its inputs and comparisons before applying it to other pharmacological datasets.⁵