EVERY MAJOR AI LAB JUST GOT A REPORT CARD AND THE HIGHEST GRADE ANYONE EARNED WAS A C-PLUS The Future of Life Institute released its 2026 H1 AI Safety Index on July 7, grading nine of the world’s largest AI companies on how seriously they are managing the risks of increasingly powerful systems. The results are not flattering. Anthropic earned the top score — a C+. OpenAI came in at C. Google DeepMind managed a C. Meta received a D+. xAI, DeepSeek, and Mistral all received failing grades. The index evaluates companies across six categories including pre-deployment testing, existential safety planning, and governance structures. The weakest area across all companies was existential safety, where no company scored above a C-minus. Most scored a D or below. The most damning finding is what happened to the commitments these labs made when they were trying to win regulatory goodwill. Companies have quietly withdrawn or softened the red line commitments that once promised development would halt if systems approached dangerous capability thresholds. Policies prohibiting military use have been successively rolled back. The internal guardrails are being lowered at exactly the moment capabilities are accelerating. The labs will continue saying safety is a priority. The independent panel says the evidence points the other way. At the current rate of capability development, the gap between what these systems can do and what infrastructure exists to manage them is widening, not closing. Keywords: AI safety rankings 2026, Future of Life Institute AI index, Anthropic OpenAI safety grades, AI red lines military use