Cyber risk is an important consideration in today’s risk management and insurance industries. However, the statistical features of cyber risk, including concerns of solvency for cyber insurance providers, are still emerging. This study investigates the dynamics of ransomware severity, specifically focusing on different statistical dimensions of extortion payments from ransomware attacks across various ransomware strains and/or variants. Our results indicate that extortion payments are not identically distributed across ransomware strains/variants, and thus violate necessary assumptions for solvency determinations using classical ruin theory. These findings emphasize the importance of re-examining these assumptions under empirical data and implementing dynamic cyber risk modelling for portfolio losses from extortion payments from ransomware attacks. Additionally, such findings suggest that removing coverage for extortion payments from insurance policies may protect cyber insurance firms from insolvency, as well as create a potential deterrence effect against ransomware threat actors due to lack of extortion payment from victims. Our work has implications for insurance regulators, policymakers, and national security advisors focused on the financial impact of extortion payments from ransomware attacks.