Anthropic has officially implemented what is being described as the strictest Claude account ban mechanism in its history, alongside the release of its 2026 usage policy. The updated policy refines penetration-style review standards for regional access, defines access restrictions for three categories of entities, adds constraints covering multiple risk scenarios, establishes a tiered handling framework, tightens regional control verification rules, and closes various loopholes used to circumvent controls. This policy tightening comes during the company's IPO roadshow stage, helping to consolidate compliance risk management and weed out problematic customers. During the same period, OpenAI also released a new version of its usage policy, set to take effect on November 20, which introduces a pre-approval requirement for model access intermediaries and cracks down on gray-market arbitrage activities such as unauthorized relay distribution.