Case Study: MetaVoice accelerates AI customer service training with Oracle Cloud Infrastructure

MetaVoice transforms phone-based customer service with Oracle AI infrastructure

MetaVoice is building human-grade AI customer service representatives to automate phone operations at massive scale, but needed fast, reliable GPU compute clusters to train its real-time multimodal AI without the burden of managing its own infrastructure. The company turned to Oracle Cloud Infrastructure, specifically OCI H100 GPU clusters and high-speed RDMA networking, to support its growth and speed up model training.

Oracle implemented OCI H100 GPU clusters with a high-speed RDMA network between bare metal instances, delivering 50% better cost efficiency than other cloud providers and enabling 3 Tbps communication across the cluster. As a result, MetaVoice trained models faster and reduced time to market, while also tripling storage to 55 TB per node to simplify data streaming and cut egress-fee complexity.

MetaVoice

Siddharth Sharma

Chief Executive Officer