Allstate explores quantum computing for insurance portfolios

Details

Allstate and IBM have published joint research (arXiv, May 2026) demonstrating how quantum-enabled methods could improve home insurance portfolio construction. The work frames the problem as a variant of the classical computer science knapsack problem: selecting the optimal combination of insurance policies to maximize value without exceeding acceptable risk thresholds — a computation that becomes intractable for classical computers at scale. The core challenge in home insurance is correlated risk. Unlike car insurance, where accidents are largely independent events, natural disasters — wildfires, hurricanes, hailstorms — affect many policyholders simultaneously. Allstate's Chief Analytics Officer Eric Huls noted this requires portfolio-level thinking, not individual policy assessment. When risks are highly correlated, modeling worst-case scenarios across thousands of policies simultaneously is exactly the kind of combinatorial problem where quantum computing may offer practical advantage. The research positions IBM's quantum work in a business context ahead of full fault-tolerant quantum systems. Allstate data scientists Jean Utke and Technical Director staff at Allstate's Insurance Product organization contributed to the work. IBM Research is building a library of applied quantum use cases spanning insurance, logistics, chemistry, and financial optimization.