Antioch Raises $32M to Bring Cloud Simulation to Physical AI Testing

When robotics companies test new features through repeated physical trials, each experiment consumes expensive equipment and engineering time, the iteration speed of physical AI cannot keep up with pure software domains. In September 2026, simulation platform Antioch announced the completion of a $32 million Series A funding round led by Greylock. Amazon's Ring has become a customer, stating that simulation results are 'highly consistent' with real physical testing. This 16-month-old New York-based startup is shifting robot development from 'physical trial and error' to 'cloud-based verification'.

For robot developers, physical environment testing has always been a critical bottleneck. Verifying new capabilities, model updates, or hardware changes requires equipment, space, and engineering time, which severely slows down development and limits test coverage. Antioch co-founder Harry Mellsop noted that robot teams often spend weeks setting up warehouses and invest millions of dollars in building test facilities. Companies like Tesla and Waymo spend hundreds of millions on simulation infrastructure each year to minimize real-world testing needs. The gap between simulation and reality lies in making virtual environments realistic enough for robots trained within them to operate reliably in the real world. Antioch builds on technologies such as NVIDIA's physics engines and world models, developing proprietary domain libraries that enable developers to simulate a vast range of test scenarios beyond what physical conditions could offer.

Antioch's platform allows engineering teams to create digital twins of hardware and run high-fidelity simulations on cloud infrastructure. A team can expand a single physical experiment into thousands of parallel simulation evaluations, identifying system weaknesses and verifying improvements before deployment. The technology stack integrates NVIDIA Omniverse, Isaac Sim, and Isaac Lab frameworks, with Nebius providing cloud infrastructure. Mellsop likens it to 'Cursor for physical AI': just as Cursor simplifies access to AI models, Antioch provides a unified simulation interface, eliminating the need for teams to rebuild workflows. The focus is on sensors and perception systems, which are core requirements for autonomous driving and drones.

Antioch's early customers include Amazon's Ring and several Fortune 500 technology and logistics companies. Ring's vice president of software, Jason Mitura, said that Antioch's simulations match physical tests closely, even in scenarios deliberately left uncalibrated, and this confidence allows them to move more testing and development into simulation, accelerating engineering cycles. Greylock partner Saam Motamedi believes physical AI and robotics are 'reaching a critical inflection point' and that Antioch is building the core development platform for the next phase. Co-founder Alex Langshur emphasized that over the past 40 years, the only economically viable path to revitalizing industrialization lies in robotics and automation, and scalable testing capability is the critical bottleneck limiting progress. As simulation platforms bring 'software speed' to the physical world, the robotics industry is approaching a day when engineers write code, run thousands of virtual validations, and confidently deploy to real machines.