Safeworld emerges from stealth with $12M+ seed for genAI robot safety testing

Safeworld, a startup focused on safety evaluation for generative AI-powered robots, emerged from stealth on Oct 5, 2026, announcing a seed round of more than $12 million. The round was led by Shine Capital and a16z Speedrun, with participation from Box Group, Carnegie Mellon University Endowment, Innovation Endeavors and SV Angel.

The company was co-founded by Dr. Ding Zhao, who directs the Safe AI lab at Carnegie Mellon University, along with veteran startup executive Kyle Wong and machine learning engineer Simo Rachidi. Zhao has spent much of his career working on AI safety. The founders argue that generative AI architectures are less predictable than traditional algorithms, and that deploying robots safely requires both advanced probabilistic evaluations and a way to build trust.

Safeworld evaluates robotic control systems in simulated environments populated with realistic human models. The company builds digital versions of real-world scenes in models like Genesis or MuJoCo, connects the robot's actual software, and runs thousands of scenarios in which virtual human models encounter the machine. Tests cover situations such as blind corners in factories, stopping distances, whether a robot detects a person carrying boxes, and tripping or falling scenarios.

The startup believes robot makers will want a third party to validate safety work, even to share information about safety cases among competitors. Gritt Robotics CTO Vishal Dugar, whose company is partnering with Safeworld, said formal mathematical proofs are difficult for these systems and safety necessarily has to be established empirically, given the variety of human postures and appearances on job sites.

Safeworld is still deciding between offering a platform for external users or a services-based approach. Ding Zhao said any company planning to deploy robots will need to pay for safety evaluation.