XtalPi pitches crystal structure prediction platform for faster solid-form selection

By AI, Created 5:22 AM UTC, May 29, 2026, /AGP/ – XtalPi is promoting its XtalGazer crystal structure prediction platform as a way to speed up polymorph screening, reduce material use and lower the risk of late-stage form surprises in pharma and materials development. The company says the workflow can cut timelines by 25% to 50% and help identify thermodynamically stable crystal forms before costly experiments begin.

Why it matters: - Crystal form selection can change a drug or material’s solubility, stability, dissolution rate, bioavailability, conductivity and mechanical performance. - Late-stage form conversion can trigger recalls, development delays and financial losses. - Traditional experimental screening can miss the most thermodynamically stable form.

What happened: - XtalPi outlined its XtalGazer™ crystal structure prediction platform as a compute-first workflow for solid-state R&D. - The platform uses first-principles quantum physics and AI-driven search to predict thermodynamically stable crystal structures. - XtalPi said the service is designed to cover polymorphs, salts, cocrystals, hydrates and solvates. - The company pointed to the platform website at the CSP service page and the main site. - XtalPi also linked to a case study on nirmatrelvir solid forms.

The details: - The workflow starts with either a 2D molecular file, such as a SMILES string or MOL file, or milligram API powder samples if the structure is unknown. - Users define the target system, such as polymorph screening or salt screening with a specific counter-ion list. - XtalPi recommends setting the search space, including space groups, temperature and pressure conditions, with support for exploration across up to 230 space groups. - The search phase uses global search algorithms to generate candidate crystal packings. - DFT energy calculations and geometry optimization provide lattice energies used to rank stability. - Routine polymorph projects are listed at 2-3 weeks, while more complex multi-component or flexible-molecule projects are listed at 6-8 weeks. - The output is a crystal energy landscape that plots lattice energy against crystal density. - Low-energy clusters are treated as plausible polymorphs, and energy differences of just a few kJ/mol can determine the predicted stability order. - Predicted XRPD patterns are compared with experimental patterns to validate known forms. - XtalPi said its MicroED service can determine crystal structures from sub-milligram microcrystalline powder when SCXRD is not feasible, with turnaround as fast as one day. - The platform is also positioned for phase-diagram modeling, form-conversion risk assessment, crystallization-condition selection and IP planning. - XtalPi said the platform can help predict intrinsic solubility, crystal morphology and mechanical properties once a structure is known. - The company said the service can improve project delivery speed by 25% to 50% compared with traditional or less optimized computational approaches. - XtalPi said a CSP study does not require physical API samples during the prediction phase and can reduce the need for dedicated experimental screening facilities. - XtalPi cited a 98% prediction accuracy rate for covering crystalline forms obtainable through experiments. - The company said its technology ranked among the top two performers out of 28 global teams in the 7th CSP Blind Test organized by the Cambridge Crystallographic Data Centre.

Between the lines: - The pitch reflects a broader shift from trial-and-error screening to a compute-first, validate-later model. - The value proposition is not just speed. It is also earlier risk detection, especially when energy gaps are small and multiple forms may compete. - The platform is being framed as useful both for discovery of new forms and for confirming that an existing form is the most stable one. - The claimed performance metrics are strong, but they are company-reported and should be weighed against project-specific validation needs.

What’s next: - Researchers using the platform would move from prediction to experimental confirmation, especially for low-energy candidate structures. - Teams are expected to use the results to guide crystallization conditions, storage decisions and manufacturing risk controls. - Early structural insight may also inform patent strategy before a solid form is widely developed.

The bottom line: - XtalPi is betting that better crystal prediction can reduce expensive surprises in solid-form development and shorten the path from molecular design to usable product.