Or build two 2.5MM DCs (if can parallelize your workload well enough) and in case of disaster, you only lose capacity.

You need however plan for 1MM+ pa in OPEX because good SREs ain’t cheap (or hardware guys building and maintaining machines)