2026-08-21 · Source: Insurance Journal
Summary in 3 Points • NVIDIA views AI compute as an investable infrastructure asset • AI Agency Insurable Value (AIV) measures economic exposure from AI systems • Autonomous systems require distinct insurance risk categories --- NVIDIA has positioned AI compute as an investable infrastructure asset, highlighting its potential to attract over $500 billion in third-party capital through partnerships with major financial entities like Apollo and Goldman Sachs. This shift indicates a broader transition where AI is seen as productive infrastructure rather than just a software expense. The concept of Agent Insurable Value (AIV) emerges as a framework to assess the economic exposure when organisations delegate authority to AI systems. AIV differentiates between the economic value of AI infrastructure and the autonomous actions of AI agents, which can significantly impact economic activities. Autonomous systems introduce unique insurance risks, such as Autonomous Resource Allocation Risk, that require distinct measurement systems.
The London insurance market could see new opportunities and challenges as AI systems become more integral to business operations. Underwriters may need to develop new frameworks to assess the risks associated with AI agents, particularly those with significant economic authority. This could lead to the creation of specialised insurance products that address unique AI-native exposures, such as Autonomous Economic Decision Risk and Autonomous Operational Failure Risk. As AI infrastructure becomes more investable, insurers might also consider the implications for policy wording, pricing, and exposure management in relation to AI-driven economic activities.