Each Cloud Server has a reserved amount of processing power on the server where it's running.
The more vCPU (or CPU Power) a Cloud Server has, the more processing cycles and power it has. Note that the Cloud Server runs on a platform that shares the processing power of some high-end machines, and the virtualization system uses the vCPUs to share the processor between tenants, and this is done dynamically.
One of the upsides of this kind of platform is that the machines can "burst" when the machine load is low. This means that at any given moment, if not all users are using the processing power, all the resources are redirected to the active users, increasing their speed momentarily. In the worst case scenario, when everybody is actively using the CPU, the system guarantees a minimum of processing cycles based on the quantity of vCPU each server have, making the best use possible of the machines, increasing their utilization levels and reducing the amount of power needed to run the servers.