Google Cloud adds an Anthropic-managed way to run Claude models

Google Cloud announced general availability of a new, Anthropic-managed deployment option for Claude models, alongside its existing Google-managed option. Both deliver the same Claude models on day zero and the same pricing and Google Cloud billing, but differ in who acts as the data processor and when new features arrive.

In the new Anthropic-managed option, Anthropic runs the stack on Google Cloud infrastructure under its own terms of service, while customers still authenticate through Google Cloud IAM, keep Cloud Audit Logging, and bill through their existing Google Cloud account. This option gets day-zero access to new Claude Platform features such as Claude Managed Agents (in beta), skills, and code execution.

The existing Google-managed option keeps Google Cloud as the data processor, with VPC Service Controls, customer-managed encryption keys, and data residency in the US, EU, and APAC, plus native integration with services like BigQuery and Model Armor. New Claude Platform features roll out to that option over time as they're integrated. Both options give access to Claude Fable 5.1, Sonnet 5.5, Haiku 5.5, and Opus 5.5.

Why it matters

Enterprises now get to choose between faster feature access or stricter governance when running Claude on Google Cloud, without changing pricing or billing.

Source: Google Cloud