Both your sources show that only 30-35% of nodes run on AWS. They seem to support my point that Ethereum would continue running during an AWS outage, not your point that Ethereum runs "solely" on AWS.

Furthermore, the cointelegraph article misquotes a Messari report as if it's discussing the current state of Ethereum as having central points of failure, when it's really discussing a hypothetical upgrade to Ethereum that was never considered nor built.

Finally, node percentages don't actually matter in terms of chain uptime. What really matters are staking validator percentages, since they are what determine chain building and finality.