1. OpenADMET receives funding for open toxicity data1
OpenADMET has announced a Gates Foundation grant to build open toxicity resources around the hERG potassium channel and aminergic G-protein-coupled receptors. The programme plans standardised functional assays, compound-bound structures, public datasets, baseline machine-learning models and prospective blind challenges. For researchers developing small molecules, the intended contribution is a shared basis for studying off-target liabilities across chemical series. The announcement describes work to be done, rather than an already released toxicity benchmark. It does not disclose the grant amount, delivery dates or evidence that the planned models improve clinical safety.1
2. Quotient reports an interim test of AI-guided formulation in humans2
Quotient Sciences says an AI algorithm reached a preset pharmacokinetic target within three dosing periods in an ongoing study with healthy participants. After each period, the system used dissolution and human exposure data to propose the next modified-release composition and dose; a safety committee approved each composition before dosing. The study uses an established generic drug to test the method, rather than develop a new product. This is a consequential company-reported step toward adaptive formulation development, but full data are promised only after completion later this year. The announcement does not establish comparative clinical efficiency.2
3. Altis closes $25 million to expand AI endpoints for oncology trials3
Altis Labs has announced the closing of a US$25 million Series A co-led by OrbiMed and Qiming Venture Partners USA. It plans to extend its imaging-based AI models to more cancer types and expand deployment with biopharmaceutical partners. Its IPRO system derives survival predictions from routine radiology images, aiming to support earlier decisions in drug trials. The financing release also cites an earlier post-hoc analysis of the MARIPOSA trial; that is supporting company-reported context, not a new prospective validation result disclosed here. The new development is capital for expanding this approach to clinical evidence generation.3
4. Polyphron announces $20 million for tissue-based AI validation4
Polyphron has announced $20 million in seed financing led by Quiet Capital for a platform combining automated human-tissue production with AI simulation. In its September 18 press release, the company says cardiac tissue is in production across multiple donor lines and that liver tissue is planned this year. The approach aims to test computational hypotheses against living systems and model how responses vary across donors. Citybiz reported the financing on September 17.5 These are company-reported capabilities and plans; the release supplies no independent benchmark showing that the platform predicts clinical outcomes.4
5. NSF opens an X-Labs route for AI in physical systems6
The U.S. National Science Foundation has opened an X-Labs funding topic for AI in physical systems, including robotics and technologies supporting scientific discovery. Its September 16 announcement also previews a separate sequence-to-function topic aimed at designing proteins and genomes with predictable biological behaviour. That future topic is not yet accepting proposals. The developments give interdisciplinary teams an immediate funding route and advance notice for assembling biological-engineering collaborations. They sit within the $1.5 billion X-Labs initiative announced in May; September’s news concerns topic availability, rather than a new funding total or completed research.6
References
OpenADMET. Putting the "T" in ADMET: Tackling Toxicity with New Funding. OpenADMET; 2026; Original responsible-organisation announcement or released report. Accessed 2026-09-20T16:16:43.840Z.
Source evidence and access
What we are building; original index dated 18 Sep 2026
Evidence paraphrase: Gates Foundation grant received for hERG and aminergic GPCR work; planned standardized experiments, structures, datasets, baseline ML and blind prospective challenges. Amount and release timetable absent.
Public responsible original announcement and linked report read; dates checked on original dated page or OpenADMET index.
Andrew Lewis; Quotient Sciences. Quotient Sciences Proprietary AI-Enhanced Solution Optimizes Formulation in Clinical Study. Quotient Sciences; 2026; Original responsible-organisation announcement or released report. Accessed 2026-09-20T16:16:43.840Z.
Source evidence and access
Dated lead; paragraphs on model retraining, safety committee and ongoing trial
Evidence paraphrase: Sep17 interim announcement; target reached in three dosing periods. Model retrained on dissolution and PK after each period; safety committee approves compositions. Established generic used for method test, ongoing dosing, full data expected year end.
Public responsible original announcement and linked report read; dates checked on original dated page or OpenADMET index.
Altis Labs. Altis Labs Raises US$25 Million Series A Co-Led by OrbiMed and Qiming Venture Partners USA to Accelerate Clinical Development with AI Endpoints. Altis Labs; 2026; Original responsible-organisation announcement or released report. Accessed 2026-09-20T16:16:43.840Z.
Source evidence and access
Announcement paragraphs 1–2; IPRO and MARIPOSA paragraphs
Evidence paraphrase: Sep17 closing US$25m SeriesA led by OrbiMed and Qiming US; planned cancer-type and deployment expansion. IPRO survival predictions from scans. Release cites earlier WCLC post-hoc MARIPOSA result; no claim of prospective validation.
Public responsible original announcement and linked report read; dates checked on original dated page or OpenADMET index.
Polyphron. Polyphron Raises $20M to Build Physical Verification Layer for AI-Driven Biology. Polyphron press release distributed by HPCwire; 2026; Issuer press release; financing also reported September17. Accessed 2026-09-20T16:22:49.754Z.
Source evidence and access
Release dateline and paragraphs1,3–5; source attribution at end
Evidence paraphrase: $20m seed led Quiet Capital; AI tissue manufacturing and simulation. Cardiac tissue across donor lines in production; liver planned this year. These are company claims, not independent performance validation.
Full public issuer-supplied release read, explicitly attributed Source: Polyphron; company site inaccessible. Earliest located reporting September17 is distinguished from dated September18 release.
citybiz. Polyphron Raises $20 Million to Build AI Verification Platform for Drug Development. citybiz; 2026; Dated business-news report; used only as evidence of reporting chronology. Accessed 2026-09-20T16:28:42.092Z.
Source evidence and access
Headline and September17 date, opening paragraph
Evidence paraphrase: page dated September17 2026 reports Polyphron seed financing of$20m.
Full publicly available original citybiz page inspected. This supports only that the story was reported September17; consequential financing/platform claims remain attributed to the issuer release.
U.S. National Science Foundation. NSF announces 3 additional topics as part of the NSF X-Labs initiative to pursue generational breakthrough science and technology efforts. NSF News; 2026; Official funding topic announcement. Accessed 2026-09-20T16:22:49.754Z.
Source evidence and access
Dated introduction; AI for Physical Systems; Future NSF X-Labs topics
Evidence paraphrase: AI for Physical Systems proposals invited, including robotics/scientific discovery; future Sequence to Function and Computation at Limit of Physics topics not yet accepting proposals. $1.5bn initiative was initiated in May, not new award today.
Full official dated announcement read; recovered missed NSF venue within seven-day window, retaining original September16 date.
Publication record
Published 2026-09-20.
Sources, selection and claims were checked in an independent AI editorial review, followed by the AI editor's approval. This is not academic peer review.
- 2026-09-20 · Version c881ea4d · Viewing this version