Datacenters are having fewer, but bigger failures

"“This is the lowest level recorded since 2020 and continues a multi-year trend of improving reliability.”"

"“Higher rack densities, load variability, and operating closer to available power limits may increase the likelihood of cascading failures,” Uptime warns."

"“This is believed to have contributed to several failures and incidents at some datacenters,” the report reads."

"“While power issues accounted for 45 percent of respondents' most impactful outages in 2025, this is down from 54 percent in 2024,” the analysts write."

Datacenter uptime has improved over the past five years, with the lowest recorded level of impactful or serious outages since 2020. Half of surveyed operators reported an impactful or serious outage in the past three years, continuing a multi-year trend of improving reliability. Failure rates are declining more slowly, suggesting diminishing returns from existing resiliency efforts and making it harder to add additional “9s” to SLAs. Analysts attribute this to greater system complexity and more challenging operating environments, including higher rack densities, load variability, and operation near available power limits that can increase cascading failure risk. Shortages of generators, switchgear, transformers, and power and cooling systems have led some operators to use second-hand or unproven hardware, contributing to failures. Power-related failures remain the leading cause of major disruptions, though their share is decreasing.

Read at theregister