Insilico Medicine – Weekly Recap

Insilico Medicine delivered a busy week of updates spanning clinical, platform and conference activities, underscoring its push to position generative AI at the core of drug discovery and development. The company highlighted new data and presentations across respiratory, gastrointestinal and oncology-focused programs, as well as advances in AI tools for chemistry.

In respiratory disease, Insilico plans four presentations at the American Thoracic Society International Conference 2026 in Orlando featuring its AI-driven inhaled Rentosertib (ISM018_055) program for pulmonary fibrosis. The company was selected as the 2026 Respiratory Innovation Summit Featured Success Story, emphasizing AI-generated TNIK inhibition targets, inhaled delivery strategies and Phase I design.

Insilico framed the inhaled Rentosertib work as part of its broader strategy to apply generative AI to high-need pulmonary indications. While no efficacy data or detailed timelines were disclosed, the focus on safety, tolerability and targeted lung exposure suggests progress toward or within early-stage clinical evaluation and could support future partnering discussions.

In gastroenterology, Insilico is presenting first-in-human Phase I results for Garutadustat (ISM5411), a gut-restricted prolyl hydroxylase domain inhibitor for inflammatory bowel disease, at Digestive Disease Week 2026. Developed via its Pharma.AI platform, the candidate’s Phase I data are being shared in a poster session and are intended to showcase advancement of AI-designed assets into human trials.

The Garutadustat update highlights Insilico’s effort to build an AI-driven IBD pipeline and to validate its platform through early clinical milestones. Favorable Phase I outcomes could support out-licensing, co-development or progression to Phase II, although the company has not yet disclosed detailed results or downstream plans.

In oncology and targeted protein degradation, Insilico emphasized its work in PROTAC drug discovery following the U.S. FDA approval of Arvinas and Pfizer’s VEPPANU. Using its Chemistry42 platform, the company reported designing a bifunctional PROTAC, D16-M1P2, targeting PKMYT1 in CCNE1-amplified cancers, with preclinical data published in Nature Communications.

The PROTAC program reportedly involved AI-driven generation and filtering of more than 2,000 molecules with iterative linker optimization, aiming to shorten traditional design–make–test–analyze cycles. This activity positions Insilico in a now validated modality and may enhance its appeal as a discovery partner in precision oncology, though the work remains preclinical.

On the AI tooling front, Insilico announced acceptance of a paper at ICML 2026 describing ChemCensor, a new benchmark for large language model-based retrosynthesis within its MMAI Gym Suite. The framework is intended to provide chemistry-aware, multi-solution evaluation and is backed by results from its C3LM model, with supporting materials slated for release on open science platforms.

If adopted, ChemCensor could improve confidence in AI-driven retrosynthesis tools and strengthen Insilico’s influence over benchmarking standards in computational chemistry. The company notes that a newer C3LM iteration reportedly surpasses several established retrosynthesis models, suggesting ongoing internal model enhancement.

Insilico also highlighted its upcoming presence at the Bio-IT World Conference & Expo in Boston, where Petrina Kamya, Ph.D., is scheduled to speak. The firm is using the event to deepen engagement with the discovery and precision medicine community, potentially supporting business development and long-term partnership opportunities.

Overall, the week’s developments show Insilico Medicine advancing AI-designed assets into the clinic, expanding its role in emerging modalities like PROTACs and investing in benchmarking infrastructure for AI chemistry, while maintaining active visibility at major scientific and industry conferences.