XtalPi launches AI platform for siRNA drug design

XtalPi unveiled Kodexia, an AI-powered platform it says can improve siRNA sequence design, chemical modification, delivery, and translation from lab results to in vivo performance. The company is targeting key bottlenecks in RNA therapeutics development, including extrahepatic delivery and freedom-to-operate risks.

Why it matters: - siRNA therapeutics can silence disease genes, but development still faces major barriers in sequence optimization, delivery, intellectual property, and the gap between cell-based and animal results. - XtalPi says Kodexia is designed to address those bottlenecks in one workflow, which could speed candidate selection and widen the range of diseases siRNA medicines can reach. - The platform is aimed at improving the odds of finding differentiated candidates with first-in-class or best-in-class potential.

What happened: - XtalPi (HKEX:2228), based in Boston, introduced Kodexia on August 24, 2026. - The company described Kodexia as an AI-centric siRNA therapeutics platform that combines generative AI with first-principles modeling. - XtalPi said the platform covers sequence design, chemical modification recommendation, experimental validation, delivery optimization and IP-informed design. - More information is available in the company's announcement.

The details: - Kodexia generates and prioritizes candidate siRNA sequences for a specific target. - The platform evaluates candidates across silencing activity, stability and off-target risk. - XtalPi said Kodexia jointly optimizes sequence design and chemical modification instead of relying on a fixed modification template. - The system recommends sequence-dependent modification patterns for each unmodified sequence. - XtalPi said internal head-to-head studies showed Kodexia-recommended modification patterns achieved stronger silencing activity than the Advanced ESC strategy in more than 70% of tests under low-concentration conditions. - The platform supported early identification of multiple lead molecules with activity comparable to or better than industry positive controls, according to XtalPi. - Kodexia is designed to support freedom-to-operate assessments by exploring sequence space beyond existing patent-protected designs. - The platform also includes in vivo efficacy prediction models to help rank candidates earlier in development. - XtalPi said those models are built on large volumes of in vivo siRNA data. - The modeling stack is tailored to hydrodynamic injection mouse models, AAV-based mouse models, transgenic mouse models and non-human primate studies. - The company said the models are intended to prioritize candidates that may outperform positive controls in initial in vivo screening and shorten the path to preclinical candidate nomination. - Kodexia also includes a delivery portfolio built from XtalPi's work in lipid nanoparticles, antibodies and peptides. - Current delivery approaches include lipid nanoparticles, small-molecule conjugation, peptide conjugation and fatty-acid conjugation. - Priority programs focus on kidney, adipose tissue and muscle delivery. - The platform includes a dual-target design workflow that treats a dual-target construct as a single chemical entity. - XtalPi said the workflow jointly optimizes the two antisense strands, linker topology and physicochemical properties from the earliest stage of design. - The company said the approach can deliver at least a two-fold improvement in R&D speed compared with conventional sequential methods. - Kodexia also uses human multi-omics datasets and AI prediction models for virtual screening of novel targets.

Between the lines: - XtalPi is positioning Kodexia as more than a sequence-generation tool. - The broader pitch is that connecting design, modification, validation and delivery could reduce the trial-and-error cycle that slows siRNA programs. - The emphasis on extrahepatic delivery signals where the field still sees the hardest unsolved problem after GalNAc-based liver delivery. - The focus on IP-aware design also suggests XtalPi is trying to help partners avoid crowded patent zones earlier in development.

What's next: - XtalPi is likely to use Kodexia to support additional internal and partnered siRNA programs across extrahepatic, dual-target and novel-target discovery. - The company said the platform is intended to improve translational success and accelerate preclinical candidate nomination. - The main test will be whether the AI-guided workflow continues to translate into stronger in vivo performance and differentiated assets.