Making operations google’s problem is precisely what drives BigQuery, FaaS and Dataflow adoption. Its operationally safer when you dont need a hadoop team to maintain your data infrastructure.

See Nytimes, Twitter and Spotify for examples.

Now if you’re talking about future-safety there was an argument...until a few hours ago.

Again, you’re talking about what’s “operationally safer” not what’s “reputationally safer”. If AWS goes down, no one is going to question your decision - and you’re in the same boat as a lot of other people. If GCP goes down and everyone else is up, people are going to ask a million question.

Every company you named isn’t the same thing as just arguing to the powers that be - AWS is what Netflix uses. AWS has many more “referencable clients” that matter.