Insider Brief
- Inner Logic raised $11.5 million in seed funding to build simulation infrastructure for testing medical devices and autonomous surgical systems across thousands of virtual patients before clinical use.
- General Catalyst and Bison Ventures co-led the round, with participation from J2VP, Defined, PTX, Valia Ventures, Page One, Alumni Ventures and Flare.
- The company said its platform uses real clinical and procedural data to simulate variations in anatomy, imaging conditions and failure scenarios, with its team already demonstrating AI systems trained in simulation on physical hardware and autonomous surgical procedures in animal models.
Inner Logic has $11.5 million in seed funding to develop simulation infrastructure that lets medical-device companies design and test products across thousands of virtual patients before they are used clinically.
According to the Maryland-based startup formerly known as Semaphor, General Catalyst and Bison Ventures co-led the round with participation from J2VP, Defined, PTX, Valia Ventures, Page One, Alumni Ventures and Flare.
Inner Logic is targeting the development process for procedural devices, including products such as heart valves and catheters, as well as more advanced robotic systems that could eventually perform portions of surgical procedures autonomously.
“Surgery is on the same path as autonomous driving, which became real once the industry built the simulation and data systems to show a car could drive itself,” co-founder and CEO Dr. Tito Porras, MD, said in the announcement. “At Inner Logic, we are building that layer for all of procedural medicine starting with the way devices are designed and tested. When device makers can develop and prove intelligent systems this way, the quality of a patient’s care stops depending on which hospital they can access or which surgeon is on call.”
Much of that development currently relies on physical prototypes tested in cadaver labs or animal studies. Those experiments can be expensive and expose a device to only a limited range of patient anatomy and conditions, according to Inner Logic.
The company is moving more of that work into computer simulation and its platform creates virtual patients using real clinical and procedural data that allows developers to test devices across thousands of simulated patients and procedures before moving into physical testing.
For conventional medical devices, the approach is intended to let engineers evaluate how a design performs across a wider range of patient anatomy than would typically be practical through physical experiments alone. For autonomous surgical systems, Inner Logic said simulation can expose AI models to variations in anatomy, medical imaging and potential failure scenarios that would be difficult to reproduce comprehensively in physical testing.
The company views that testing infrastructure as a step toward greater surgical autonomy, in which robotic systems can perform parts of a procedure with less direct intervention from a surgeon.
Inner Logic said research by its team has already demonstrated the approach in orthopedic trauma, where AI algorithms were trained entirely in simulation before being deployed on physical hardware and patient anatomy.
The team has also worked on autonomous robotic surgery. According to the company, a system developed by its researchers completed critical steps of gallbladder removal in a live animal model without intervention from a surgeon.
Inner Logic was co-founded by Porras, who left a neurosurgery residency at Johns Hopkins Hospital to start the company; Chief Technology Officer Mathias Unberath, a computer-vision and simulation researcher at the Johns Hopkins Whiting School of Engineering; and Chief Robotics Officer Axel Krieger, an associate professor at Johns Hopkins.
Krieger’s research has included a robot that performed autonomous laparoscopic surgery in a live animal, according to Inner Logic. The company said its three founders have collectively authored more than 400 peer-reviewed publications and are named inventors on more than 40 patents.
