The first cluster, the OCI Zettascale10, will be hosted in the company’s OCI public cloud. It will enable customers to configure AI environments with up to 800,000 Nvidia graphics processing units.