Executive-grade visibility into how the hyperscalers really perform. We track and benchmark major customer-impacting outages across AWS, Azure and Google Cloud - then attribute root cause and model the impact.
Corpus last updated June 30, 2026
A documented methodology, an open dataset, and a clear point of view. The work that informs every figure across the network.
A modeled snapshot of the three hyperscalers - uptime, incident volume and root cause. These figures are illustrative, not live measurements; for the measured public-incident corpus, see Research & Data →
Illustrative - modeled figures, not live measurements
across AWS · Azure · GCP
11.8 hrs of impact
3-provider weighted mean
down 9% vs 2025
99.96%
rolling 12-month uptime
47
incidents YTD
214m
downtime YTD
99.94%
rolling 12-month uptime
61
incidents YTD
318m
downtime YTD
99.97%
rolling 12-month uptime
39
incidents YTD
176m
downtime YTD
Entra ID auth degradation across West Europe
Global auth · multi-region
us-east-1 DynamoDB elevated error rates
us-east-1 · API throttling
Cloud Load Balancing config rollback
us-central1 · networking
Storage account latency spike, East US 2
East US 2 · storage
An illustrative full-year model - sample actuals through June with a projected second half.
Monthly customer-impacting downtime per provider with a model-projected second half, plus the quarter-over-quarter picture across AWS, Azure and Google Cloud.
Shaded region (Jul–Dec) is model-projected from trailing run-rate.
Stacked customer-impacting minutes, all providers.
Every figure we publish traces to a named primary source - a provider post-mortem or status page - corroborated by the reporting of established newsrooms and analysts.
“We don’t ask you to take our word for it. Each claim links to the source that stands behind it.”
CloudDowntime Research · Editorial standard
Briefings, benchmarks and playbooks - published as the data moves.
Slack’s incident history - from the 2021 new-year outage to routine degradations - and what it reveals about running always-on collaboration at scale.
How OpenAI’s API and ChatGPT have held up under explosive demand - the recurring capacity and degradation patterns, and what they mean for teams building on it.
Anthropic’s reliability posture for the Claude API - status transparency, degradation patterns, and how it compares as a production LLM dependency.
Track the hyperscalers the way the analysts do. Start with the 2026 annual report.