Case Study: Swiss Re achieves scalable, cost-efficient internal risk modeling on Azure with MathWorks

Swiss Re Migrates Risk Model to the Cloud with MATLAB Parallel Server

Swiss Re, a global reinsurance leader, faced rapidly growing compute demand from its FINMA‑approved internal capital adequacy risk model (millions of simulations, +120 data packages) while operating a private‑cloud cluster that constrained scalability, performance consistency, and cloud migration plans. To move to Azure and satisfy internal security and operational policies, Swiss Re partnered with MathWorks to adopt and adapt MATLAB Parallel Server and related deployment artifacts for a secure, cloud‑ready reference architecture.

MathWorks collaborated with Swiss Re to customize the Azure Marketplace MATLAB Parallel Server image and deployment (private VNETs, no public endpoints, amended post‑deployment scripts), while Swiss Re rearchitected parallelization, added VM scale sets with programmatic scaling, and built a DevOps image pipeline. The combined solution delivered measurable benefits: on‑demand scaling cut average usage by ~23% despite a +128% rise in annual calculation demand, converted a fixed 40‑worker pool into ~3,000 independent jobs for greater resilience and throughput, and supported high‑capacity clusters (e.g., 282 CPUs, 5.1 TB RAM) with faster VM types yielding up to +93% per‑worker speed—resulting in more efficient, available, and cost‑effective risk calculations.

Swiss Re