In the seven years since its IPO, Medicilon has helped bring nearly 700 IND applications to clinical approval — around 100 a year, which works out to one new drug crossing the IND threshold and entering the clinic on Medicilon’s platform every four days or so.
Racing timelines on one side, tightening global compliance on the other: speed and quality control look like natural adversaries. How does Medicilon step outside the industry trap of “fast means fragile, strict means slow” and build a distinctive R&D advantage?
One-stop collaboration: removing the waiting losses between stages
Traditionally, medicinal chemistry, CMC, pharmacology and efficacy, pharmacokinetics, safety evaluation, and IND filing are carried out serially by separate providers. Technology transfer, divergent standards, and information asymmetry generate heavy communication and workflow losses, and the breakpoints in that chain become a leading cause of delay.
Medicilon’s one-stop platform instead brings drug discovery, CMC, preclinical studies, and IND filing under a single project-management and quality-standard system. At kick-off, scientists, technical leads, and the quality and regulatory teams all engage upfront and in parallel to lock down R&D endpoints, global regulatory requirements, and key delivery milestones. Modules are then planned and advanced in parallel, so data come from one source, standards stay unified, and work flows internally — with no technical gaps between vendors.
Running in parallel does not mean simplifying the science. Protocols, endpoints, and QC standards are aligned with NMPA and FDA regulatory logic from the design stage; for China–US dual filings and multi-region submissions, dual compliance tracks are built in from the outset, so nothing has to be torn down and rebuilt later over a regulatory conflict. The chain from molecular discovery to IND filing is connected end to end.
AI-assisted drug discovery: shortening iteration at the source
Drug discovery is where timelines and failure risk concentrate. The traditional route relies on high-throughput wet-lab trial and error, with slow iteration across molecular design, synthesis and screening, and developability assessment — and large amounts of time spent on molecules that were never viable.
Medicilon has built the MAIDD (Medicilon AI-Driven Drug Discovery) platform around four AI sub-platforms spanning small molecules, nucleic acids, biological macromolecules, and peptides. AI supports a dry-wet development loop — model building, virtual screening, molecular design (generation), wet-lab validation, and data feedback for optimization — raising both efficiency and quality. The same AI capabilities are applied to PROTAC, ADC, peptide/cyclic peptide, and small nucleic acid platforms for the construction and optimization of linkers, sequences, and other modules, strengthening the platform’s capabilities and lead. The loop keeps feeding data back into the models, accelerating candidate discovery and lead optimization.
Identifying risk earlier: cutting the cost of late failure
Drug development fails often, and much of the lost time comes from pursuing the wrong direction, choosing the wrong model, or discovering a developability problem too late. Some projects only reveal their flaws at the GLP toxicology stage, when everything invested before is sunk.
Drawing on more than two decades of drug R&D experience, Medicilon has built a library of over 1,100 pharmacodynamic models and pairs it with frontier evaluation technologies — AI-based ADMET prediction, organoids, and NAMs (new approach methodologies) — to move risk identification substantially earlier. At the early screening stage, absorption, metabolism, toxicity, and off-target risks are anticipated in advance, improving candidate developability and the odds of clinical translation. For frontier modalities — the ones above plus CGT and GLP-1 — mature impurity studies, formulation strategies, and bioanalytical packages can be reused quickly and adapted to each project. What this compress is redundant experimentation; the rigorous evidence chain that supports an IND filing is kept intact.
An international compliance system: reducing variables in filing and review
At the IND stage, the most expensive item is rarely the experiment itself. It is data that are not accepted, repeat studies that are demanded, or work redone because standards were misaligned. Medicilon was among the earliest companies in China to hold both NMPA and OECD GLP accreditation while meeting FDA GLP requirements, and both its Shanghai and Boston R&D centers hold AAALAC accreditation. China–US, China–Australia, and even multi-country filing requirements can be designed into protocols from the start, so domestic studies do not have to be repeated against overseas standards. By the end of June 2026, the company had helped 77 China–US dual filings and 7 China–US–Australia triple filings reach clinical approval.
Quality follows the QbD principle, with control embedded across the whole process: the quality team takes part in protocol review, and raw records, sample transfers, instrument calibration, and pathology reads are traceable end to end. Digital systems — Provantis, EMPOWER, Chromeleon, and LIMS — together with SEND conversion and eCTD filing capability keep data verifiable, internationally recognized, and directly usable in a submission.
Efficiency comes from speed, and rests on rigor. As new therapeutic modalities keep emerging and global regulatory science keeps evolving, Medicilon will keep strengthening its one-stop preclinical R&D platform, its international GLP compliance system, and its frontier technologies — holding a differentiated position between quality and efficiency, and moving more innovative molecules from Target to IND.
