Mantis Biotech Builds Physics-Based Digital Twins of Athletes from Multimodal Data

Mantis Biotech, a New York startup that raised a $7.4 million seed round led by Decibel VC with participation from Y Combinator, a few angel investors and Liquid 2, builds physics-based digital twins of the human body, predictive models of anatomy, physiology and behavior. TechCrunch reported in March 2026 that one of its main clients is an NBA team. "We're able to take all these disparate data sources and then turn them into predictive models for how people are going to perform," said founder and CEO Georgia Witchel. The company targets a gap in biomedical AI: models trained on structured healthcare data struggle with edge cases such as rare diseases, where representative data is scarce and ethical and regulatory limits restrict the use of patient records. The platform ingests textbooks, motion-capture video, biometric sensors, training logs and medical imaging. An LLM-based system routes, validates and synthesizes these streams, and a physics engine turns the combined record into high-fidelity renders that become synthetic training data for predictive models. The physics layer grounds the synthetic data in realistic anatomy and lets the team generate datasets that do not exist publicly; Witchel's example is hand-pose data for a person missing a finger, produced by removing the finger from the model and regenerating it. For the NBA client, Mantis builds representations of each athlete that track how jumps change day by day over a year against factors such as sleep and arm load. The seed money is going to hiring and go-to-market. Mantis plans to release the platform publicly for preventive healthcare and is working with pharmaceutical labs and FDA-trial researchers on modeling patient responses to treatment.