Safeguarded AI aims to provide quantitative safety guarantees for AI. Their current funding round is for demonstrations that AI systems with such guarantees are useful and profitable in safety-critical contexts (e.g. optimising energy networks, clinical trials, or telecommunications).