Tariff tensions undermine trust in cloud hyperscalers

Over the past decade, hyperscaler cloud providers have earned their customers’ trust — an achievement they can take pride in. Many large enterprises and small startups now rely on them to host mission-critical workloads. Prices have generally fallen over time, and while there have been outages, most were relatively short-lived (see Outage data shows cloud apps must be design for failure). Crucially, there have been no significant security breaches within the hyperscalers’ infrastructure. With the exception of some AI-backed instances, customers have historically had ready access to resources, enabling them to spontaneously scale their applications with changing demand.

This trust has been established through sustained performance rather than through financial assurances. Service level agreements mean little in practice — the compensation offered during an outage is minimal compared with the actual loss incurred. Moreover, compensation is only paid out on the failed services, not the whole application. There is also no compensation for security breaches, poor performance, or other metrics (see Cloud SLAs punish, not compensate).