Mass General Brigham’s Sriraman Says SaaS Outages Expose Hidden Cloud Dependencies

In this Health System CIO Show episode, Mass General Brigham CTO Sri Sriraman explains how recent public‑cloud outages exposed a blind spot in many health‑system SaaS contracts. Providers often deploy services across multiple hyperscaler regions, yet their control planes remain tightly coupled, so a failure in one region can cascade to another. Sriraman’s open letter to SaaS vendors demanded clear regional segregation and transparency, underscoring that hidden cloud dependencies pose real patient‑safety risks and jeopardize business continuity.

Sriraman also outlines Mass General Brigham’s evolving technology strategy. While the $23 billion organization plans to migrate most commodity workloads to the cloud for cost efficiency, it retains on‑premise capacity for heavy GPU‑intensive AI research. He warns that AI subscription models, like OpenAI’s token‑based pricing, can quickly inflate expenses without strict governance. By negotiating enterprise rates and measuring ROI, the health system balances productivity gains against the potential for runaway spend.

For CIOs facing similar challenges, Sriraman recommends a disciplined resilience questionnaire for every SaaS partner. Ask about regional hosting, control‑plane isolation, fail‑over times, and who initiates recovery. This proactive approach restores the on‑premise habit of scrutinizing disaster‑recovery designs, ensuring vendors disclose inter‑region interconnects and DR readiness. By demanding these details now, health systems can avoid surprise outages—like the benefits‑portal failure on a critical deadline—and protect both patient safety and operational budgets.