@RookAils1
As far as I understand, it operates on the FCFS principle (First come, First serve).
In the scenario below, I have a "testflow" that I executed locally and simultaneously, I assume a scheduled flow is invoking the same flow or another flow on the same machine. However, Power Automate Desktop prioritizes the flow that was called first (either local attended or cloud attended).
For instance, let's assume the local attended invoked "testflow" at 5:43:10 AM, and the Cloud attended invoked "testflow" at 5:43:11 AM. In this case, Power Automate Desktop prioritizes the local attended invocation, and the cloud attended invocation may be queued.
If both invocations occur at the exact same second, like 5:43:10 AM, it may consider microseconds, although such scenarios are very rare at the moment.
In simpler terms, it's like making web service calls where the server responds to the first request it receives, utilizing our virtual machine to execute our bot functionality. That's essentially how it works.

Hope this helps!!
If I've resolved your query, please consider accepting it as the solution and giving it a thumbs up to help others find answers faster.