SafeWorld Raises $12.2 Million to Build Robot Safety Testing Platform Using AI Simulation

SafeWorld, a Palo Alto-based AI lab spun out of Carnegie Mellon University's Safe AI Lab, emerged from stealth on October 5, 2026 with $12.2 million in seed funding to build simulation-based safety validation tools for autonomous robots.

The seed round was co-led by Shine Capital and a16z Speedrun. Additional investors include Box Group, the Carnegie Mellon University Endowment, Innovation Endeavors, and SV Angel, alongside angel investors and executives from NVIDIA, Google DeepMind, Waymo, Meta, and Salesforce.

The company was co-founded by Dr. Ding Zhao, director of the Safe AI Lab at Carnegie Mellon and a former researcher at Google DeepMind; Kyle Wong, CEO and former head of the Stanford startup accelerator StartX; and Simo Rachidi, a repeat founder and former principal security and machine learning engineer at Salesforce Einstein.

SafeWorld's platform uses high-fidelity 3D simulation tools, compatible with physics models like Genesis and MuJoCo, to test how robots behave in rare, dangerous, or unexpected edge-case scenarios. The system runs thousands of simulations involving reactive human models, helping companies evaluate the safety of their robotic systems before physical deployment.

The startup is addressing a gap that has grown as generative AI allows robots to operate with greater autonomy. Traditional hard-coded safety measures are no longer sufficient when robots can make decisions on their own. SafeWorld provides third-party empirical safety reports that serve as a trust layer for robot builders and operators facing liability or regulatory scrutiny.

The company is currently working with Gritt Robotics, which builds AI-driven robots for industrial-scale solar farm construction, and is running early pilots in automotive manufacturing, warehouse automation, and medical device production.