> from AWS or MS about their serverless offerings and other such lock in mechanisms

IME in Cybersecurity/Enterprise SaaS, F1000s largely eschewed serverless capabilities and control plane lock-in features (eg. Fargate, Autopilot) because they wanted to be able to reduce single vendor risks as well as negotiate better contracts.

Most firms I've dealt with that size tend to have at least 2 hyperscalers used internally, plus an on-prem footprint that is increasingly being reduced.

The firms I've seen use serverless and lock-in features the most tended to be smaller shops (eg. Mid-markets, hyper scaling startups) that simply don't have the bandwidth to invest in a large DevSecOps org but are also technical enough to not get locked into an MSSP contract.

Also, OCI is on an absolute warpath right now - a LOT of very large tech-first F500s are getting and signing OCI sweetheart deals as we speak. They've replicated GCP's GTM approach, which is ironic because Thomas Kurian is ex-Oracle and was the de facto "keeps the lights on" guy there, until Catz and Ellison pushed him out.

Interesting. So, the largest serverless installation that I'm aware of is one that has a massive number of in the field sensors posting data every five minutes, think many millions of sensors. They're in a complete panic there. And they're not exactly small, rather the opposite.