Claude arrives on Google Cloud: Everything you need to know about the new model managed by Anthropic
If you are working with large language models (LLMs) in your organization, you are surely familiar with the dilemma: on one hand, the desire to use Anthropic's most advanced technology, and on the other hand, the need to maintain a secure, managed work environment integrated into your existing cloud infrastructure. The good news is that this integration has become much simpler and more accessible.
Anthropic recently announced the General Availability of Claude Models on Google Cloud (Anthropic managed). This is a significant step that allows developers and organizations to consume Claude models directly through the Google Cloud environment, while maintaining maximum flexibility [1, 6].
What has actually changed?
Until now, working with Claude within Google Cloud was done primarily through Google-managed services. Now, Anthropic is offering an additional path where it manages the technological stack itself. This means you get access to all the capabilities of the Claude platform (such as Claude Managed Agents, code execution capabilities, and advanced tools) on launch day, while using the identity (IAM) and billing of your existing Google Cloud account [1, 2, 5].
Why is this important for organizations?
The key difference between the Google-managed path and the Anthropic-managed path lies in the question of who processes your data. For organizations with strict regulatory, privacy, and information security requirements, this choice is critical [1, 4]:
- •Identity and security management: You continue to use Google Cloud IAM to manage permissions, which allows for seamless integration with your organizational security policy [1, 5].
- •Access to innovation: In the Anthropic-managed path, you get immediate (Day-Zero) access to the newest models, beta features, and research, without needing to wait for updates to the Vertex AI platform [1, 4].
- •Unified billing: There is no need to open new accounts or deal with additional vendors; everything is included under your existing Google Cloud account [2, 4].
How to choose between the paths?
To help you get organized, here is a quick comparison:
| Feature | Google-managed | Anthropic-managed | | :--- | :--- | :--- | | Data processor | Google Cloud | Anthropic | | Feature availability | According to the Vertex AI schedule | Immediate access (Day-Zero) | | API interface | Adapted for Agent Platform | Compatible with the original Claude API |
If your organization requires Google Cloud standards as the data processor with support for VPC Service Controls or CMEK, the Google-managed path remains the preferred choice [1, 2]. Conversely, if you want the full Claude experience with all the newest features, while utilizing Google's infrastructure for management and billing, the Anthropic-managed path is the ideal solution for you [2, 4].
Summary and conclusions
This move by Anthropic and Google is further evidence that the cloud is becoming the natural home of enterprise artificial intelligence. The ability to choose who manages the infrastructure behind the model allows developers to build smarter applications, faster, without compromising on the organization's security requirements.
If you are already working in a Google Cloud environment, I recommend checking out the new models in the Model Garden and starting to experiment with the new Claude Platform. This is the time to check how the new features can improve your workflows.
For more information: