Using a confidential KMS allows us to support elaborate confidential inferencing services composed of several micro-services, and products that need multiple nodes for inferencing. For example, an audio transcription https://ai-studio-seven.vercel.app