gcp Google Cloud Blog ·

SeaVerse reduces infrastructure costs by 60% with GKE Agent Sandbox

blogaigcpgapreviewarchitectgcp-bigquerygcp-gke
announcement

SeaVerse, a gaming startup building a platform for playable AI experiences, adopted Google Kubernetes Engine (GKE) and GKE Agent Sandbox. This implementation reduced their infrastructure costs by up to 60% while providing strong isolation, better observability, and flexible resource allocation for dynamic, multi-tenant AI workloads. The GKE Agent Sandbox leverages Kata Containers+Cloudhypervisor or gVisor for kernel-level isolation, supporting persistent file systems and fast allocation rates at general availability. This foundation streamlines the creative loop for users, enabling immediate and easily shareable AI experiences with integrated logging and monitoring.

Notes (1)
  • SeaVerse leverages GKE Agent Sandbox for playable AI experiences

    SeaVerse utilized GKE and GKE Agent Sandbox to build a platform for playable AI experiences, allowing users to create and interact with AI-generated content. This infrastructure choice addressed challenges with multi-tenant sandbox workloads, providing strong kernel-level isolation, improved observability, and cost flexibility by supporting both microVM and gVisor isolation runtimes. The solution helps ensure immediate and easily shareable AI experiences, with native logging and monitoring integrated into sandboxed environments.

Read the original announcement →

https://cloud.google.com/blog/products/containers-kubernetes/seaverse-chooses-gke-agent-sandbox/

Related releases