Valgo is a public benefit corporation building a risk quantification layer for physical AI, with a focus on autonomous vehicles and other safety-critical systems. Its work helps inform the validation, deployment, and insurance of autonomous technologies by making risks measurable and safety assessments more rigorous. Valgo combines research with practical tools, including Human Crash Baselines, which calculates human crash-rate comparisons for robotaxis and autonomous trucks and makes its methodological choices transparent.

The company’s research spans safety validation, failure probability estimation, rare-event and failure-scenario discovery, risk assessment, and planning under uncertainty. Its published work includes algorithms and frameworks for testing black-box systems, identifying potential failures, and evaluating autonomous vehicle safety, alongside research on machine learning certification and aircraft collision avoidance. Valgo also shares its findings through research papers, white papers, and educational resources, including a textbook on safety validation algorithms. By translating advanced methods into tools and evidence for assessing risk, Valgo aims to support informed decisions about physical AI systems.