Google Cloud Now Lets Anthropic Run Claude

Google Cloud made an Anthropic-managed deployment of Claude generally available on October 23. It runs on Google Cloud infrastructure and bills through Google Cloud, but Anthropic operates the stack and acts as the data processor.

For builders, this is a choice between feature speed and compliance posture. Same models, same price, different processor, and a different date on which new platform features show up.

Image: Google Cloud Blog, Expanding Claude availability and deployment choice on Google Cloud, October 23, 2026.

Two deployments, one bill

The new Anthropic-managed option gets Claude Platform features on day zero. Google names Claude Managed Agents (in beta), skills, and code execution. You still authenticate with Google Cloud IAM, keep Cloud Audit Logging, and draw down eligible Google Cloud commitments. The terms are Anthropic's Commercial Terms.

The existing Google-managed option stays. Google is the data processor under your existing Google Cloud agreement. It offers VPC Service Controls, customer-managed encryption keys, and data residency in the US, EU, and APAC, plus integrations such as BigQuery and Model Armor. New Claude Platform features arrive there once Google integrates them.

Both options get new Claude models on day zero at the same pricing. Google lists Claude Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5 as available.

The decision is legal, not technical

Read the comparison table closely. Google's post lists VPC Service Controls, CMEK, and regional residency under the Google-managed option. It does not claim them for the Anthropic-managed one. Do not assume parity until Anthropic's documentation says so.

Changing the data processor is a procurement event. Your Google Cloud agreement may already cover Claude traffic today. Moving to the Anthropic-managed option puts that traffic under a different vendor's terms, and your DPA review should treat it that way.

The upside is real for agent teams. Managed Agents, skills, and code execution have so far been the reason some teams called Anthropic's API directly instead of going through their cloud. This option keeps those features inside Google Cloud identity and billing.

What to do this week

Split workloads by requirement. Anything that needs VPC Service Controls, CMEK, or pinned residency stays on Google-managed. Agent workloads that need Managed Agents or code execution now are candidates for Anthropic-managed.

Get legal and security sign-off before the first call, not after. Tag spend by deployment option so commitment drawdown stays traceable. And document which option each service uses, because the feature sets will drift as betas launch.

Sources

https://cloud.google.com/blog/products/ai-machine-learning/expanding-claude-availability-on-google-cloud