Yes this is one of the most powerful capabilities that MongoDB Atlas allows. You can change a cluster from single to multi-region, or change from one region to another, or change the cloud provider of a cluster all without downtime: this is made possible by making these changes in a rolling manner across the distributed system while maintaining majority quorum.
(Note for completeness that there will be a momentary replica-set level election which takes a few seconds, which drivers by default deal with gracefully by taking advantage of retryable writes.
And separately note that network access to a cluster in one region or cloud provider may not be possible from another region or cloud provider without some special care, especially if you’re taking advantage of in-region private networking like Private Endpoints or VPC peering. One option would be to take advantage of public IP-level access on the IP Access List at least during the transition period.)