At my place it's not the cloud provider that is the issue, or cloud itself. It's the governance and control of access which cripples the usage of cloud and damages innovation, exploration, change, improvements, ideas. It's a red tape issue and a lack of willingness to give up control and to trust IT pros to do their jobs.
My favourite memory for calls with people (as the owner of one of these cloud platforms):
* a project screamed they had to have a call, turns out that their selected container/service provider must have full admin of the entire AWS org to work. My response was short: No. The service was national infrastructure and the folks, the devops needed counter terrorism checks (CTC), plus their software would be incompatible... etc..
* a project data scientist wanted to take production PII data onto a laptop for processing during a late security architecture review. Kicking and screaming resulted - always funny when a project doesn't do its work, gets the sign off from legal, risk and security. Don't shoot the person that points it out. It makes me sound like a BOFH.. but when your platform holds the entire group of airlines, subsidiary service companies and airport systems and your calendar has three meetings on the report submission, and then covering the report in steerco with CDIOs, CISCO, Head of Risk, CTO, and SLT from multiple companies.. You damn right I'm gonna make you the professional you ain't.
I created a sandbox system, outside of the production/internal connectivity, that allowed teams to operate as they saw fit. The sandbox was charged to the cost code too, their budget owner had to give it the thumbs up (all automated approvals) but they could do what they wanted (within the platform permissions) which included external access to market place and AI etc.
I had personal (restricted cost and session time) that was paid for by the global budget - the idea was for people to use for AWS learning. You wouldn't believe the number of projects that attempted to get me change the charge restrictions so they could charge outside of their budget codes! No public internet for you on these either!
This made it quick for PoCs, but it's also safe and secure for people learning. The Dev environments were stricter and aligned with production, so you'd not be developing something that would be incompatible in the real environments.
All our platform code was in a repo for anyone in the company to view. It meant they could see what the system would do (and how best to circumvent we found on one instance!).
The result accelerated use and transformation. It became a standard way of working. They didn't have to spend time integrating with the BSS/OSS/security/networks, the CMDB or finops - it was all automated and integrated into the organisation. The CISO switched from NO to happy the baselines the platform provided, it meant that security review meetings were faster.