Clinical Intelligence
AI-discovered and computationally designed drugs in clinical trials — tracked in real time.
Showing 15 of 15 trials
Positive Phase 2a results published in Nature Medicine (June 2025)
First fully AI-discovered drug to show Phase 2a efficacy. 60 mg QD arm showed +98.4 mL mean FVC improvement vs decline in placebo. Published in Nature Medicine (IF 58.7).
PandaOmics (target ID), Chemistry42 (molecule generation), InClinico (clinical prediction)
Pivotal study initiated mid-2025; SABCS 2025 subset analysis presented
10.3-month median PFS overall, 11.0 months in 2L. 33% ORR across all patients, 53% in kinase-mutant subset. First mutant-selective PI3Kα inhibitor — avoids hyperglycaemia of pan-PI3K drugs.
Dynamo™ (motion-based drug design using MD simulations + ML)
Positive Phase 1b/2 TUPELO trial data (Dec 2025)
75% of evaluable patients showed polyp burden reduction; 43% median reduction at 12 weeks (4 mg QD). Rapid, durable response in rare genetic cancer predisposition syndrome.
Recursion OS (phenomics-based discovery from cellular imaging at scale)
First patients dosed (Dec 2025); interim data expected H2 2026
First AI-designed antibody to enter clinical trials. De novo generated using generative models — never existed in nature. Phase 2 for endometriosis planned Q4 2026.
Integrated Drug Creation™ (generative AI antibody design from zero-shot, no animal immunization)
Phase 1 results presented at ERS 2025; IPO raised $400M (Feb 2026)
AI-engineered ultra-high affinity anti-TSLP antibody with superior half-life vs comparators. Designed for 6-month dosing interval. Phase 1 showed durable TSLP suppression.
Chroma (generative protein diffusion model), proprietary ML for affinity/stability optimization
FDA-approved (Dec 2023); pediatric data (ages 5-11) presented ASH 2025
First CRISPR therapy approved by FDA. 97% of SCD patients free of vaso-occlusive crises post-treatment. Pediatric expansion (ages 5-11) showed comparable efficacy. Global regulatory submissions planned H1 2026.
Computational guide RNA design and off-target prediction algorithms
HEART-2 trial ongoing; 14 patients treated as of Mar 2025
One-time in vivo base editing of PCSK9. Well-tolerated, no treatment-related SAEs. Dose-dependent LDL-C reduction observed. Phase 2 planned for 2025. Could replace lifetime statin therapy.
Computational base editor design, LNP optimization with ML-guided formulation
FDA-approved Dec 2025; DESTINY-Breast09 pivotal
44% reduction in disease progression or death vs THP (HR 0.56). First new 1L regimen for HER2+ mBC in 12 years. Computationally optimized DXd payload and cleavable linker.
Computational linker-payload optimization, ML-based toxicity prediction
Priority Review granted (Mar 2026); Breakthrough Therapy Designation
Expansion into early breast cancer. FDA Priority Review with PDUFA date Q3 2026. Breakthrough Therapy Designation reflects substantial improvement over existing options.
Computational linker-payload optimization, predictive PK/PD modelling
Phase 1 data presented Oct 2025; dose escalation ongoing
AI-discovered selective CDK2 inhibitor with brain penetration. 28% partial response rate in non-CNS tumours, 33% in patients with measurable brain metastases. Acquired by Bristol Myers Squibb for $800M.
Enchant™ (physics-informed ML for potency/selectivity), Magnet™ (generative chemistry)
Phase 1 CTA approved (Australia, Sep 2025); dosing initiated
AI-designed PRMT5 inhibitor entering clinic. XtalPi's physics-based AI predicted crystal polymorphs and binding modes, accelerating lead optimization. Phase 1 in advanced solid tumours.
XtalPi ID4 platform (quantum mechanics + ML for crystal structure & binding prediction)
Phase 1/2 enrolling; first AI-designed precision oncology molecule
AI-designed CDK7 inhibitor with built-in biomarker strategy for patient selection. Showcases Exscientia's precision design approach — molecule and companion diagnostic co-designed from the start.
Exscientia Centaur Chemist™ (active learning + generative design), patient-selection AI
Phase 3 HAELO trial enrolling; Breakthrough Therapy Designation
In vivo CRISPR gene editing — no cell extraction needed. Phase 1/2 showed 95% mean reduction in HAE attacks. Near-complete elimination of breakthrough attacks. One-time IV infusion.
Computational guide RNA design, LNP formulation optimization, off-target scoring algorithms
Phase 1 initiated 2025; first wholly-owned clinical candidate
First clinical candidate from Schrödinger's own pipeline (not partnered). MALT1 protease inhibitor designed entirely with physics-based computational methods. FEP+ predicted binding affinities within 1 kcal/mol.
FEP+ (free energy perturbation), WaterMap, physics-based molecular design platform
Phase 2 data expected 2026; AI-identified novel target-indication pair
AI identified PDE10A as a novel target for ulcerative colitis — a connection not found by traditional methods. Knowledge graph mined biomedical literature to surface non-obvious target-disease links. Partnered with AstraZeneca.
Benevolent Platform™ (knowledge graph + NLP for target identification, repositioning)
Data sourced from ClinicalTrials.gov, company press releases, and conference presentations. Last curated April 2026.