From what I remember, EMRs - particularly parts that do things like manage blood banks and medication dispensers - aren't just something you can have a team of consultants from Accenture vibecode, or even plancode. In several countries, they fall under the same regulations as medical devices and are subject to the same scrutiny.
I wouldn't want to be the hospital executive sitting for a deposition on a medical malpractice suit, explaining how instead of using Epic or Cerner or whomever, they decided to let AI and a bunch of recent college grads from the lowest bidder consulting firm replace a known system. Sounds like a good way to wipe out whatever you saved in costs with court judgments.
Also, switching EMRs is a huge pain in the ass. When I was a fresh-faced employee at an EMR company they sent me and other employees out to help deploy a new system in a client's hospitals in another city. This took a small army of employees, contractors, travel nurses, and consultants to do. Your ass was up at 3 AM, back at your hotel room at 8 PM. Nurses didn't care about what your program did, they wanted it a certain way and they wanted it fixed now. You're hopefully not going to have the hospital leadership saying, "Yeah, you can try this and if you fail, we'll switch again in three years". I can't imagine many healthcare systems doing that, particularly if the physicians are a major component of management.
All of what you are saying is absolutely true, but frankly doesn't matter at the executive level.
If it is a board priority to extract favorable terms from vendors (and it absolutely is right now), we will get it done consequences be damned. If you can't do it, we'll fire you and replace you with someone else. You saw this with enterprises making 12-18 month roadmaps to completely tear out VMware ESXi and migrate to Nutanix.
Unlike Broadcom which has a much more diversified business and purchased actual market leaders which allows them to be so vicious, Oracle's SaaS products have a much weaker hand as the headline of churn is much more destabilizing for a market laggard like Cerner or NetSuite than choosing to drop from 90% gross margins to 40% gross for strategic customers - and purchasers know that.
As such, as a business who is not in a position to protect against strongarming purchaser you need to preemptively build additional margins slack where possible, and it is in this vein that the NetSuite and Cerner layoffs happened today.