Mantis Biotech Builds AI ‘Digital Twins’ to Solve Healthcare Data Gaps

Mantis Biotech is developing a platform that uses synthetic data and physics-based modeling to create “digital twins” of the human body, aiming to address gaps in biomedical data that limit how AI systems perform in areas like rare diseases and edge-case conditions.

The New York–based startup says its system aggregates information from sources including medical imaging, biometric sensors, motion capture, and training logs, then uses a large language model to organize and validate that data before running it through a physics engine. The result is a high-fidelity simulation of human anatomy and behavior that can be used to train predictive models.

CEO Georgia Witchel said the approach is designed to generate datasets where real-world examples are scarce. “We’re able to take all these disparate data sources and then turn them into predictive models for how people are going to perform,” she said, pointing to applications that range from modeling athletic performance to forecasting medical outcomes.

The physics layer plays a central role in the system’s design. By grounding synthetic data in realistic representations of human movement and structure, the platform can simulate scenarios that would be difficult or impossible to capture in real datasets. Witchel described how the model could generate training data for uncommon conditions, such as hand movements involving missing fingers, by modifying its underlying simulations.

Mantis is positioning these digital twins as tools for research, training, and clinical experimentation. Potential uses include testing new medical procedures, supporting surgical robotics, and analyzing how patients might respond to treatments. The company is also exploring applications in professional sports, where it has already begun working with teams to track and model athlete performance over time.

The startup recently raised $7.4 million in seed funding led by Decibel VC, with participation from Y Combinator and other investors. The capital will support hiring and go-to-market efforts as Mantis continues building out its platform.

Witchel said the longer-term goal is to expand into preventative healthcare and pharmaceutical research, including use cases tied to FDA trials. The company is also considering a broader release of its technology, though it remains focused on refining its models and expanding its data capabilities in the near term.

The effort reflects a growing push to overcome one of AI’s biggest limitations in healthcare: the lack of high-quality, representative data in complex or uncommon cases. By generating synthetic datasets anchored in physical modeling, Mantis is attempting to create a more flexible foundation for prediction and experimentation across the biomedical field.

This analysis is based on reporting from TechCrunch.

Image courtesy of Mantis.

This article was generated with AI assistance and reviewed for accuracy and quality.